Seasonal Diet and Habitat Preferences of Cattle (Bos Indicus ...

\\SEASONAL DIET AND HABITAT PREFERENCES OF CATTLE (Bos indicus), KONGONI (Alcephalus buselaphus)

AND WILDEBEEST (Connochaetes taurinus) GRAZINGON A COMMON RANGE.

A THESIS SUBMITTED IN PARTIAL

FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

MASTER OF SCIENCE

IN

RANGE MANAGEMENT

IN THE FACULTY OF AGRICULTURE

OF THE

UNIVERSITY OF NAIROBI.

pv

W. EGO

1996.

DECLARATION

THIS IS MY ORIGINAL WORK AND HAS NOT BEEN PRESENTED FOR A

DEGREE IN ANY OTHER UNIVERSITY.

SIGNATURE

THIS THESIS HAS BEEN SUBMITTED FOR EXAMINATION WITH MY

APPROVAL AS UNIVERSITY SUPERVISOR.

SIGNATURE __MR. D. M. MBUVI

DATE f b — ^ ._______

THIS THESIS HAS BEEN SUBMITTED FOR EXAMINATION WITH MY

APPROVAL AS UNIVERSITY SUPERVISOR.

SIGNATURE ____________________________________PROF. C .N . KARUE

/(o / V- I ^ (qDATE

DEDICATION

To my parents who thought I could do better than to look

after cattle, brothers, sisters and grandmother and all

who have encouraged me and contributed to my academic

achievements since 1970 and to my grandfather, who dearly

loved me, but could not live long to see this thesis.

Special dedication to my wife, Damaris, my children

Jerotich and Kokwo and to all those who cherish the beauty of the African bush.

ACKNOWLEDGEMENTS

I wish to thank the European Economic Commission

(EEC) for providing the scholarship, which enabled me to

pursue this study. I am very grateful to the Director

Kenya Agricultural Research Institute (KARI) for granting

me study leave during the course of the study. I am also

very grateful to the owner of the Game Ranching Ltd., Dr.

D. Hopcraft, and Dr. M. Sommeralte, the Director of

Research for allowing me easy access to the Ranch and to

use their camping facilities. The Ranch manager Mr. P.

Tilley and N. Nyamu were co-operative and friendly.

I express my sincere gratitude to Mr. D.M. Mbuvi and

Prof. C.N. Karue, under whose supervision this study was

carried out. Their patience, untiring guidance, helpful

suggestions and constructive criticisms, despite their

busy schedules, are highly appreciated. Special thanks

to the Director, National Range Research Centre, Mr.

P.F.K. Kibet, for allowing me to use laboratory

facilities and other resources. Thanks to Dr. H.K.

Cheruiyot of KARI for fruitful discussions,

encouragement, comments and friendship. I am greatly

indebted to the late Dr. T. Tandingar and Mr. F. Nang'ayo

for their assistance in data analysis.

I thank Mr. J.M. Kinyua and S.M. Mwangi for their

great assistance both in the field and laboratory work.

Their sense of humour made the ticks and thorns bearable.

IV

ABSTRACT

The feeding habits of cattle (Bos indicus), kongoni

(Alcephalus buselaphus) and wildebeest (Connochaetes

taurinus) whether as single species or in combination

with other animal species on the same range have been

studied by several researchers in different environments.

Generally these studies conclude that animals do select

their diets from an of array plants depending on what is

available to them and the prevailing conditions.

This thesis is the result of a study conducted at

Game Ranching Ltd. situated at the Athi Kapiti Plains,

Kenya, between January and August 1993, to determine the

diet and habitat preferences of cattle, kongoni and

wildebeest. The following six vegetation types

(habitats) occur in the ranch; Themeda grassland,

Balanites glabra tree grassland, Balanites - Acacia tree

grassland, Acacia drepanolobium dwarf tree grassland,

Acacia woodland and Acacia xanthophloea bushland.

The diet preference by the three herbivores was

determined by using microhistological analysis technique.

The three dominant grasses in the ranch Themeda triandra,

Digitaria macroblephara and Penisetum mezianum also

formed the major diets of the animals. The browse

component of the diets increased during the dry season by

about 100% irrespective of the animal species, with

cattle always having twice as much browse as the wild

v

herbivores. The animals, however, selected similar diets

in terms of plant species during both seasons. Dietary

overlaps were always above 75%, but higher during the dry

season, than during the wet season. The overlaps were

lower between cattle and wild herbivores than between the

wild herbivores.

During the wet season Balanites glabra tree

grassland was the most preferred while all other habitats

had negative preference indices, with Acacia woodland and

Acacia xanthophloea bushland habitats being avoided

completely by kongoni and wildebeest. During the dry

season the two herbivores shifted their habitat

preferences to Balanites - Acacia tree grassland habitat,

with Acacia xanthophloea being avoided completely.

Balanites - Acacia grassland seem to be the best

available habitat in the ranch as it is the habitat upon

which kongoni and wildebeest depend for their survival

during the dry season. Balanites glabra grassland,

however is the most preferred habitat but is only

available during the wet season.

vi

TABLE OF CONTENTSPage

Declaration __________________________________________ ii

Dedication ___________________________________________ iii

Acknowledgements _____________________________________ iv

Abstract _____________________________________________ v

Table of contents ____________________________________ vii

List of figures ______________________________________ xList of tables _______________________________________ xi

CHAPTER ONE: GENERAL INTRODUCTION

1.1 Introduction ____________________________________ 1

1.2 Objectives ______________________________________ 3

1.3 Justification ___________________________________ 4

1.4 Hypothesis ______________________________________ 6

CHAPTER TWO: LITERATURE REVIEW

2.1 General feeding strategies of wild anddomestic ruminants _____________________________ 7

2.2 Feeding strategies of cattle, kongoni andwildebeest ______________________________________ 9

2.3 Relationship between feeding strategy andbody size _______________________________________ 11

2.4 Role of special senses in grazing _____________ 11

2.5 Influence of standing biomass on forageutilization by herbivores ______________________ 12

2.6 Forage preference by domestic and wildherbivores ______________________________________ 15

2.7 Determination of range herbivore diets ________ 20

2.7.1 Utilization technique ___________________ 20

2.7.2 Direct observation 21

vii

222.7.3 Stomach content analysis

2.7.4 Microhistological technique ____________ 24

2.7.5 Fistula technique _______________________ 27

CHAPTER THREE: STUDY AREA

3.1 Location and physiography 28

3.2 Climate 30

3.3 Animals 33

CHAPTER FOUR: MATERIALS AND METHODS

4.1 Vegetation inventory 34

4.1.1 Herb layer sampling 34

4.1.2 Woody plant density and canopy cover sampling 34

4.2 Dietary sampling 35

4.2.1. Preparation of plant reference slides 35

4.2.2 Preparation of slides from faecal material 36

4.2.3 Slide quantification 374.3 Animal density and habitat preferences 384.4 Data analysis 39

4.4.1 Seasonal diet preference 394.4.2 Dietary overlaps 414.4.3 Animal density and habitat preferences 43

CHAPTER FIVE: RESULTS AND DISCUSSIONS

5.1 Vegetation and animal diets 455.1.1 Woody plant cover and density 455.1.2 Herb layer 525.1.3 Diet composition by plant species 63

viii

715.1.4 Seasonal dietary overlap between

animals _________________________

5.1.5 Discussion ______________________________ 75

5.2 Animal density and habitat preferences ________ 81

5.2.1 Number and distribution of animals _____ 81

5.2.2 Animal density and habitat preference __ 82

5.2.3 Discussion ______________________________ 88

CHAPTER SIX: CONCLUSIONS AND RECOMMENDATIONS _______ 91

REFERENCES ___________________________________________ 93

APPENDICES 101

ix

LIST OF FIGURESFigure Page

1. Geographical location of Game Ranching Ltd. andits vegetation types ___________________________ 29

2. Monthly rainfall during the study periodcompared to the mean monthly rainfall in the study area ______________________________________ 3 2

3. Map showing the areas sampled in eachvegetation type for animals ____________________ 40

4. Prominent grass and browse in animal dietsduring the wet and dry seasons _________________ 65

5. Percent animal diet overlaps during the wetand dry seasons _________________________________ 73

6. Seasonal changes in habitat preference for kongoni and wildebeest for the six habitat types ___________________________________ 86

x

LIST OF TABLES

Table Pace

1. Mean and annual rainfall for the period 1981 to 1992 31

2 . Animal numbers and density in the study area 33

3 . Total area covered by each habitat and their respective areas that were sampled during the animal census 39

4 . Mean absolute and relative canopy cover and density of woody plants for Balanites tree grassland 45

5 . Mean absolute and relative canopy cover and density of woody plants for Balanites - Acacia tree grassland 46

6 . Mean absolute and relative canopy cover and density of woody plants for A. drepanolobium dwarf tree grassland 47

7 . Mean absolute and relative canopy cover and density of woody plants for Acacia woodland 48

8 . Mean absolute and relative canopy cover and density of woody plants for A. xanthophloea bushland 49

9 . Analysis of variance of density of trees per hectare in the study area 50

10 . Analysis of variance of percent canopy cover of trees in the study area 50

11. Mean density per hectare and percent canopy cover of trees in each habitat in the study area 51

12 . Mean density per hectare and percent canopy cover of each tree species in the study area 52

13 . Mean herbaceous standing biomass in Kg per hectare for each habitat during the wet and dry seasons 53

14 . Seasonal percent standing biomass of dominant grasses and forbs in the Themeda grassland 54

xi

55

15. Seasonal percent standing biomass of dominantgrasses and forbs in the Balanites tree grassland _____________________________________

16. Seasonal percent standing biomass of dominantgrasses and forbs in the Balanites - Acacia tree grassland __________________________________ 56

17. Seasonal percent standing biomass of dominant grasses and forbs in the A. drepanolobium dwarf tree grassland ___________________________ 57

18. Seasonal percent standing biomass of dominantgrasses and forbs in the Acacia woodland ______ 58

19. Seasonal percent standing biomass of dominant grasses and forbs in the A. xanthophloea bushland 59

20. Analysis of variance of seasonal percent standing biomass of dominant grasses and forbs 60

21. Mean percent standing biomass of dominantgrasses and forbs in each habitat ______________ 61

22. Mean seasonal percent standing biomass ofdominant grasses and forbs _____________________ 62

23. Mean percent relative densities of thebotanical composition of the diets of thestudy animals during the wet season ___________ 64

24. Mean percent relative densities of thebotanical composition of the diets of thestudy animals during the dry season ___________ 6 7

25. Mean diet preference indices of cattle,kongoni and wildebeest with percent plant availability during the wet and dry seasons ___ 68

26. Analysis of variance of mean diet preferencesof cattle, kongoni and wildebeest _____________ 69

27. Mean diet preferences of cattle, kongoni andwildebeest_____________________70

28. Percent similarity indices and correlation coefficients between cattle and wildlife diets during the wet season ____________________ 72

29. Percent similarity indices and correlation coefficients between cattle and wildlife diets during the dry season ____________________ 75

xii

30. Habitat utilization by the kongoni, showing relative densities per Km2, percentage frequencies and preference indices during thewet and dry seasons ____________________________ 84

31. Habitat utilization by the wildebeest, showing relative density per km2, percentage frequencies and preference indices during thewet and dry seasons ____________________________ 85

32. Analysis of variance of mean habitatpreferences of kongoni and wildebeest _________ 87

33. Mean habitat preferences of kongoni andwildebeest 88

xiii

CHAPTER ONEGENERAL INTRODUCTION

1.1 IntroductionKenya's rangelands are in demand for various

competing land uses due to the expanding population, eg

settlement, agriculture, pastoralism, conservation and

tourism. These land uses have caused conflicts in the

rangelands. The expansion of human population has

created the need for greater agricultural production,

forcing agricultural people to move to prime rangelands.

This has forced the traditional pastoralists to poorer

rangelands, previously occupied solely by wildlife. The

latter has been forced into small pockets and sometimes

important migratory routes have been blocked.

Governments have established Wildlife Reserves and

National Parks to facilitate tourism and to protect

wildlife. However wildlife migrates from these Reserves

and National Parks into neighbouring pastoral areas and

ranches. As a result wildlife in some areas moves to

neighbouring agricultural lands destroying crops and thus

causing friction between man and wildlife. With the

foregoing, a key issue is the future role of wildlife in

the rangelands. Further, several countries have imposed

complete restriction on consumptive use of certain

wildlife species to avoid their extinction leaving

tourism as the only legitimate use. These policies have

promoted increase of wildlife populations with resultant

1

conflicts with those of livestock producers being in the

forefront. Wildlife destroy fences, compete for food

resources and are carriers of diseases to the detriment

of livestock producers. Compensation schemes have never

been adequate for such damages.

Natural forage is the main source of feed for both

wild and domestic herbivores in the arid and semi-arid

rangelands. Pratt and Gwynne (1977) stated that grass

will remain the cheapest source of livestock feed in

Kenya in the foreseeable future due to limited supply of

grain and the cost of manufactured feed. Ruminants

(bovids) are highly efficient compared to simple

stomached herbivores (equids) in the conversion of forage

to meat (Church 1975). Duncan et al. (1990), however

argues that equids compared to similarly sized grazing

bovids have higher rates of food intake, which more than

compensates for their lesser ability to digest plant

material and are capable of extracting more nutrients

from at libitum forage diets than bovids, especially from

very high and very low fibre diets. This implies that

given an equal amount of feed, bovids would efficiently

convert it to meat than equids but the equids can extract

more nutrients from an ad libitum diet than bovids.

However, maximum livestock production is dependent upon

proper management of the resources. The most fundamental

of these is stocking the range with the correct

kinds/class and numbers of animals (Heady 1975) .

Unlike stall-fed animals that receive their rations

2

in amounts and proportions dictated by the livestock

owners, free ranging animals choose their diets from the

complex variety of available forage plant species in the

plant communities they utilize. Man therefore exerts

only limited managerial control through such decisions as

stocking rates, herd composition and size, and location

of grazing areas. The diet that an animal will

ultimately select in a particular situation and location

is a function of many interacting plant and animal

related factors (Malechek and Provenza 1983). Certain

characteristics of individual plants which influence

their acceptance or rejection therefore play a major

role. Further unique morphological, physiological and

other characteristics of a particular plant species

interact to determine the animal's feeding strategy as

they exploit the available food resources. The aggregate

effect of all these is manifested in the feeding

behaviour termed "selectivity", that is the consumption

of some plant species or plant parts to the exclusion of

others (Heady 1975, Malechek and Provenza, 1983).

1.2 ObjectivesThe objectives of this study were;

a) to determine seasonal diet preferences and

dietary overlaps of cattle, kongoni and

wildebeest, grazing on the same range and

b) to determine the seasonal habitat preferences of

the cattle, kongoni and wildebeest.

3

1.3 JustificationEast African rangelands have highly seasonal growing

conditions and periodic fluctuation of large herbivore

populations. This results from periodic shortage and

replenishment of feed resource due to seasonality of

rainfall. Foraging is an important component of fitness

and many aspects of behaviour and morphology are shaped

by the need to gather food. Foraging efficiency in part

determines the inclusive fitness of an individual, and

though food acquisition is central to activities, it

competes for time with other activities such as mating,

territory defence and predator vigilance (Krebs 1978).

As natural selection favours individuals with highest

inclusive fitness, animals are under pressure to forage

efficiently. This notion of foraging efficiency has led

to studies that investigate diet preference, competition

and habitat preference. Problems arise from the nature

of herbivore diet and constraints it imposes on the

digestive process which include those of animal

physiology which dictate the diet selection, foraging

time limitations and nutritional constraints from food

(Belovsky and Jordan 1978, Demment et al. 1986 and Van

Soest 1982).

The degree of diet selection determines the

dispersion and availability of food items for a

herbivore. Highly selective herbivores have their diets

compost of buds and flowers, which are highly digestible

but making up only a tiny fraction of the plant biomass.

4

This type of food is highly dispersed, relatively rare

and only small amounts can be harvested (Geist 1974,

Githaiga 1991). By lowering its acceptance threshold, a

larger proportion of the plant community becomes

available as food for a grazer, bites are almost

continuous and search time is reduced (Jarman 1974). The

patterning of the nutrient content and its distribution

in the vegetation is therefore critical in feeding

strategy of a ruminant grazer depending on its

selectivity regime.

Field (1968) noted that in some parts of Uganda and

East Africa as a whole, overstocking of domestic animals

by pastoralists, leads to overgrazing and range

destruction. He concluded that in many such cases the

value of wildlife is neglected and it is the first to

suffer following degradation of its habitat. Thus the

management of domestic animals do affect the survival of

wildlife. It is therefore important that the influences

of interrelated multiple use of grazing strategies should

be considered in range management to successfully meet

the objectives of finding optimal use of the range, for

livestock or for wildlife or a combination of livestock

and wildlife.

Knowledge of feeding habits and habitat preference

of both wild and domestic animals is necessary in order

to solve problems arising from the issue of preservation

and food requirement of wild animals. Information on

dietary habits of both wild and domestic herbivores is

5

therefore an important tool to a range manager in

determining what competition exists among different range

animals and in balancing wildlife and livestock numbers

with available forage (Holechek et al. 1982a, Wangoi

1984). An indication of the plant species consumed

enables a manager to know what the key plant species are

and animal performance (Holechek et al. 1981). Further

information on the feeding habits of animals utilizing a

common range is important in offering a basis for

assessing the usefulness of the range to the animals

dependent on it. Consequently data and information on

food habits and habitat preferences are important in

making management decisions, like the manipulation of the

vegetation to achieve the desired objective(s).

1.4 HypothesisThe following hypothesis were formulated in trying

to test this problem

HOi - Cattle, wildebeest and kongoni have similar

diet preferences,

H02 - Cattle, kongoni and wildebeest prefer similar

habitats irrespective of seasons.

6

CHAPTER TWOLITERATURE REVIEW

2.1 General feeding strategies of wild and domesticruminantsThe way in which ruminant species select their diet

is termed the feeding strategy (Hanley 1982). The

selectivity of each ruminant diet is expressed by; the

amount of grass versus browse, the choice of plant species

within each forage consumed and the amount of each plant

part consumed. Pregastric fermentation of ingested plant

material supplies ruminants with energy and nutrients in

the form of microbial products (Hungate 1966). In

ruminants, the initial consumer of plant primary

production is the anaerobic microbial population of the

rumen. This buffered anaerobic fermentation, supplies

nutrients to ruminants in the form of volatile fatty

acids, the end products of fermentation. The growth of

microbial cells produce proteins and vitamins which are

harvested from the rumen by passage to the abomasum where

peptic digestion occurs. This gives the advantage to the

ruminants over simple stomach mammals with hind gut

fermentation, where most microbial cells are lost in the

faeces. This advantage is especially important when plant

products, such as cell wall carbohydrates, which are not

digested by enzymes secreted by mammals, are a large

portion of ingested food. Another advantage is that plant

7

CHAPTER TWOLITERATURE REVIEW

2.1 General feeding strategies of wild and domesticruminantsThe way in which ruminant species select their diet

is termed the feeding strategy (Hanley 1982). The

selectivity of each ruminant diet is expressed by; the

amount of grass versus browse, the choice of plant species

within each forage consumed and the amount of each plant

part consumed. Pregastric fermentation of ingested plant

material supplies ruminants with energy and nutrients in

the form of microbial products (Hungate 1966). In

ruminants, the initial consumer of plant primary

production is the anaerobic microbial population of the

rumen. This buffered anaerobic fermentation, supplies

nutrients to ruminants in the form of volatile fatty

acids, the end products of fermentation. The growth of

microbial cells produce proteins and vitamins which are

harvested from the rumen by passage to the abomasum where

peptic digestion occurs. This gives the advantage to the

ruminants over simple stomach mammals with hind gut

fermentation, where most microbial cells are lost in the

faeces. This advantage is especially important when plant

products, such as cell wall carbohydrates, which are not

digested by enzymes secreted by mammals, are a large

portion of ingested food. Another advantage is that plant

7

toxins are substantially modified by the rumen

fermentation which in most cases renders them less toxic

and harmless. Since the ingested plant material is

retained in the rumen for fermentation, ruminants may be

more restricted by the fermentation characteristics of

plant material than in large simple stomached herbivores

(Van Soest 1982).

Van Soest (1982) reported that intake is the most

important parameter in the nutritional status of all

mammals. He also concluded that the rumen has a limited

capacity and rumen fill is considered an important factor

regulating intake. The turnover and capacity of the rumen

must therefore place limits to intake as it is the major

site of digestion. Welch and Smith (1969) in their

studies concluded that as plant cell wall increases in the

ruminant diet, rumination time increases to a limit at

which point intake must decrease. Rumination, which

reduces ingesta to a size that will pass through the

reticulo-omasal orifice, is therefore an important factor.

Hofmann (1973) suggested that the omasum plays an

important role in regulating rumen turnover and that its

structure and place in the digestive tract suggest a

sieving and pumping function. Therefore, the omasum also

regulate the passage rate of digesta to the lower tract.

In contrast, mammals such as zebra and elephants have no

blockage to the passage of ingesta. This gives them the

ability to increase intake in response to decreased

quality of food beyond the capabilities of ruminants.

8

These relationships suggest that ruminant herbivores need

to select diets that correspond to the limitations

inherent in the functional anatomy of the digestive tract

of each species.

2.2 Feeding strategies of cattle, kongoni and wildebeestThe prehensile capabilities of cattle, kongoni and

wildebeest is related to their feeding strategies (Jarman

and Sinclair 1979). Dentition, muzzle width and manner of

grazing or browsing are important in the ease with which

herbivores can select plants with different morphological

attributes. The most important aspect of the feeding

strategy of cattle is that they are domesticated. Their

grazing patterns and habitats in which they feed are

controlled by man. Traditional methods of keeping cattle

in bomas at night severely restrict their feeding time.

This would force cattle to be less selective when feeding

because of time limit. On the other hand, the ability of

cattle as well as sheep and goats to increase fill under

conditions of restricted feeding time (Hoppe 1977) may be

an important feature in their adaptation to domestication.

The relatively broad muzzles and efficient use of

tongues in wildebeest and cattle allow rapid and efficient

harvesting of grass leaves from dense leafy swards.

However in tall grass communities kongoni with a long,

narrow, flat muzzle is more capable of selecting for grass

leaves than cattle (Jarman and Sinclair 1979). Kongoni

are classified as "bulk feeders and roughage eaters" in

9

the subclass of the "roughage grazers" (Hofmann 1973).

Jarman (1974) described kongoni as being rather

unselective for grass species but more selective for plant

parts or growth stages. Kongoni select for maximum intake

of grass leaves in the wet season, but in the dry season

grass sheath is selected over grass stem (Stanley-Price

1977). Kongoni appear to be able to select for grass leaf

in tall grass vegetation types (Stewart and Stewart 1970,

Talbot and Talbot 1962) . Wildebeest are classified by

Hofmann (1973) as "bulk and roughage eaters" in the

subclass "fresh grass grazers dependant upon water". In

Jarman's (1974) classification scheme, wildebeest are in

the same class as kongoni, unselective for grass species

but more selective for plant parts and growth stages.

Wildebeest show preference for short grass vegetation

types (Mentis and Duke 1976). Unlike kongoni, wildebeest

do not seem capable of selecting for grass leaf in tall

grass vegetation types. Instead, they maximize intake of

grass leaves by migration or association with less

selective ungulates (especially zebra) in grazing

succession (Bell 1971). Zebra create a more acceptable

sward structure for wildebeest in the grazing succession

along the catena change from tall grass to short grass

vegetation types. This process is also repeated across

the rainfall gradients which create short and tall grass

vegetation types as occurs in the Serengeti plains (Bell 1971) .

10

2.3 Relationship between feeding strategy and body sizeA number of authors have related feeding strategies

of African ungulates to their body weight (Vessey-

Fitzgerald I960, Jarman 1974, Van Soest 1982). The

concept that very small ruminants (<15Kg) are selective

feeders and very large ruminants (>200Kg) are relatively

unselective is generally accepted. The relationship

between energy requirements, body weight and digestive

tract capacity supports this conclusion (Van Soest 1982,

Mentis 1977) . The consequence of this relationship is

very important to the feeding strategies of ruminants

differing in body weight (Van Soest 1982) . Very small

ruminants such as duikers, suni and dik-dik need to select

for diets that have both high rates of digestion and high

digestive tract passage rates to maintain rumen turnover

and intake levels necessary to meet their nutrient

requirements. Plants that meet this criteria are very low

in abundance (Jarman 1974, Mentis 1977). Consequently

these animals are restricted in specific habitats.

2.4 Role of special senses in grazingThe senses of sight, smell, taste and touch are

involved in diet selection (Arnold and Dudzinki 1978,

Arnold 1966a, Krueger et al. 1974) . Sight is the most

important in orientating the animal to other animals and

its environment. Sheep do recognize conspicuous food

plants by sight but do not use sight to help them graze

selectively. This was documented by Arnold (1966a) who

11

found that blind folded sheep ate similar diets to those

of sheep that could see under a wide range of conditions.

Arnold (1966b), in a series of studies with surgically

treated sheep to produce single and multiple sensory

impairments showed that, not only were there marked

changes in acceptability of plant species when each of the

senses were impaired, but that total food intake was

affected. He found that inability to taste had the effect

of improving the acceptability of more species than did

the inability to smell or feel them.

Arnold and Dudzinki (1978) indicated that chemical

signals influence food selection. These are received at

receptors for taste and smell. Stimuli are transmitted to

the brain and the animal responds behaviourally or

physiologically to the messages they contain (Krueger et

al. 1974). The animal then responds by integrating these

messages with others, such as feedbacks on the current

nutritional state of the animal. The desire to eat may

then result in lowering either taste or smell thresholds

of rejection (Goatcher and Church 1970, Arnold and Dudzinki 1978).

2.5 Influence of standing biomass on forage utilizationby herbivoresOne of the most important properties of the East

African rangelands is the abundance of plant biomass on a

temporal and spatial basis, which is mainly dependent

largely on rainfall. Rainfall is the major determinant of

12

the quality and quantity of the forage available for

consumption by herbivores, thereby determining the

abundance of both plant and animal components of which the

East African grasslands support a greater biomass and

diversity of herbivores than any other terrestrial

ecosystem (Van Dyne et al. 1980). This combination may

result from high degree of resource partitioning among

these herbivores, with large populations and high

diversity permitted by relatively efficient use of

available food and space (Lamprey 1963, Jarman and

Sinclair 1979).

The growth form (i.e. height, leaf/stem ratio and

crown structure) of tropical grasses has an important

effect on the eating time, bite size and intake (Stobbs

1973, Chacon and Stobbs 1976). They concluded that in

general tall growth of tropical grasses leads to a longer

eating time, smaller bite size and lower intake by cattle

when compared to short leafy growth of the same grasses.

Herbage yield is also negatively related to bite size and

intake. These relationships indicate that the presence of

large amount of grass biomass would be deleterious to

ruminant utilization. The ability of grazing ruminants to

maximize forage intake on short grass swards may be a

factor in migratory patterns and choice of vegetation

types of wild ruminants in East Africa. The grazing

succession described by Bell (1971), in which Zebra create

a short grass sward more acceptable to wildebeest and

Thomson's gazelle, fits this concept. Wildebeest also

13

migrate to the short grass plains in the Serengeti

ecosystem in their period of peak demands during

parturition and early lactation (Sinclair and Norton-

Griffiths 1979).

Concomitant with large seasonal changes in plant

standing biomass are equally large and important changes

in nutrient contents and digestibility. Although there

are infinite variations in nutritive value among plant

species, there are some similarities among plant groups.

Generally the grasses and forbs of the herb layer, have

relatively high digestibility (French 1957) and high

concentrations of nitrogen, phosphorus and other nutrients

(Bredon and Wilson 1963, Taerum 1970) soon after growth

resumes at the onset of rainy season. During the period

of early growth, concentration of crude protein in grass

leaves is approximately 8-20%, while forbs have a higher

concentration ranging from 15-30% (Dougall et al. 1964).

As the growth ages, its nutritional quality declines as a

result of increases in structural carbohydrates, so that

both nutrient concentration and digestibility decrease

(French 1957, Kilcher 1981). By contrast, browse plants

in tropical grasslands generally have deeper root system

which exploit a less ephemeral moisture supply and store

food reserves in stems and leaves rather than in roots

(Lawton 1968, Owen-Smith 1982). Browse, therefore do not

decrease much in protein and carbohydrates, with advancing

maturity as much as grasses do (Heady 1975).

14

2.6 Forage preference by domestic and wild herbivoresNumerous studies geared towards understanding the

feeding habits of both wild and domestic herbivores have

been conducted. The evidence assembled todate indicate

that ungulates are selective in their diet for at least

part of the year, involving habitat selection, plant

species selection, and selection for plant parts (Jarman

and Sinclair 1979). As diet quality declines, with higher

incidence of fibrous material, more time at the expense of

feeding must be set aside for rumination which is thus a

critical component of foraging. Thus what a grazer does

when not foraging is as important as it does in overall

feeding strategy (Demment et al. 1986). The fibrous

material content of the diet ultimately controls ingestion

and assimilation of other nutrients (Van Soest 1967). The

constraints imposed by the digestive physiology dictate

that ruminants select an easily digestible diet of high

quality. This selectivity with a preference for green

grass parts, has been documented by several studies in

domestic and wild herbivores (Sinclair 1972, Stobbs 1975,

Talbot and Talbot 1962, Gakahu 1982).

Ruminants have been found to display different

selectivity regimes closely associated with body size with

profound effects on the ecology and behaviour of the

species. Gwynne and Bell (1968) showed that larger

animals tolerate coarser, lower quality food, and smaller

ungulates can coexist with larger species by using

scattered remnants of high quality food. Larger species

15

therefore facilitate plant/forage utilization by smaller

species by removing coarser material and exposing the

higher quality portions. However reduction of resources

below a critical level can trigger competition (Field

1972) . Hillman and Hillman (1977) concluded that food

resource shortages are so intense in drought years as to

cause high mortality among ungulates. Further evidence of

resource shortage comes from observations of a shift in

diet composition among plant parts. Andere (1981)

concluded that if resource abundance varies with season

and there is evidence of seasonal shortage, then niche

overlap may be construed as actual competition.

Wildebeest and zebra are virtually pure grazers and

select leaves, which have the highest ratio of protein and

soluble carbohydrates to cellulose (Gwynne and Bell,

1968) . In the dry season there is a decrease in the

intake of leaves, at the expenses of leave sheath and

stem. Similarly Owaga (1975) found that both wildebeest

and zebra are almost pure grazers, taking little forbs

(about 1-2%) during the wet periods and almost none at

other times. The proportions in zebra diet were usually

close to the availability and therefore seemed to be

relatively random feeders. In his experiment in Kruger

National Park, South Africa, Ben-Shahar (1991) showed that

there was a considerable overlap of grass species

composition in the diets of zebra and wildebeest. However

wildebeest diet alternated with seasons, showing high

preferences during the winter for grasses which were

16

rejected during the summer.

Field (1975) noted that cattle, buffalo, eland and

oryx grazed within the grass/herb layer during the early

growth period. He also noted that annual and drought

tolerant grasses form the main diet of oryx, while buffalo

and cattle feed on bulky perennial grasses. A study of

goat and eland diets on the Kiboko Research Station, Kenya

by Ng'ethe and Box (1976) showed that the bulk of diets of

both animal species consisted of leaves from relatively

few plant species. Although elands utilized a wide

variety of plants, they consumed a larger proportion of

grasses than goats. Elands are mixed feeders. Van Zyl

(1965) reported that elands browsed 76.5% and grazed 23.5%

of their time in the field, while Lamprey (1963) concluded

that elands selected 70% grasses and 30% browse species.

Kerr et al. (1970) reported that grasses were minor forage

components for elands. These conflicting conclusions

imply that intake could be largely dictated by the

nutritional status of the animal, locality and

availability of forage. Further they confirm the

observation that elands are capable of utilizing a wide variety of plant species.

Field and Potere (1972) have documented that cattle

prefer grazing to browsing and that sheep like fine grass,

forbs and shrubs while goats are mainly browsers. In a

study on the feeding behaviour of cattle in a semi-arid

part of Tanganyika, Payne and MacFarlane (1963) noted that

cattle were browsing more frequently as the dry season

17

advanced. Holechek et al. (1982b) investigated the

seasonal diets of cattle in the Oregon forests, United

States of America. The study showed that grasses, forbs

and shrubs averaged 61%, 16% and 23% of the diet

respectively. Composition of the diet differed with

advancement of the season. Forbs were heavily used in the

early part of the growing season before maturation.

Browse comprised as much as 47% of the diet when green

grass was unavailable. They concluded that cattle were

opportunistic grazers and did not limit their selection to

grass species. Kayongo Male (1986) studying the seasonal

variability in cattle diets in Marsabit District, Kenya,

concluded that during the wet season, annual and perennial

grasses made up the bulk of cattle diets. When the dry

season become severe the herbs, dwarf shrubs, trees and

litter constituted major portion of cattle diets. This

was in general agreement with other studies by Payne and

MacFarlane (1963). Wangoi and Hansen (1987) investigating

the seasonal dietary habits of camels, cattle, sheep and

goats grazing a common range in Marsabit District, Kenya,

concluded that although cattle predominantly ate grasses,

the browse component of their diet was higher in the wet

season than in the dry season. More than 50% of the sheep

diet consisted of grasses for all except one season, when

the browse component of their diet tended to increase

during the very dry and very wet season. Goats also

tended to browse relatively more during the driest season.

However, for camels which have mouth parts adapted for

18

browsing just like sheep and goats, the grass component

was highest during the dry season.

From most of the studies done so far, there is a

general conclusion, that grazing animals do select their

diet from an array of plants depending on what is

available to them and prevailing conditions. The animals

tend to be opportunists utilizing whatever is available

thereby showing great variation in feeding habits in

different ecological regions and certain seasonal

variation of these habits within the same region. To

consider dietary data as forage classes (grass, forbs and

shrubs) overlooks in general the important fact that

animals select their diets on a plant species basis and

even plant parts. It is important therefore that dietary

selection studies consider analysis and reporting of

diets at species level. Seasonal dietary shifts can be

abrupt particularly in areas having distinct wet and dry

seasons like in the temperate areas. The nutritional

consequences of these shifts are probably great in terms

of competition among animals on a common range.

Generalization from these diet selection studies are

difficult because all have been conducted under conditions

of different plant availabilities. Consequently, the

results tend to be location specific in terms of

applicability to management. However, such studies when

applied to the site from which the data originated, can

provide range managers with some basis for making management decisions.

19

2.7 Determination of range herbivore dietsThe procedures that have been used to determine the

botanical composition of grazing animal diets include;

- utilization technique,

- direct observation of the animal(s),

- stomach content analysis,

- microhistological technique and

- fistula techniques.

2.7.1 Utilization techniqueUtilization is one of the oldest approaches used to

evaluate grazing animal's diet (Holechek et al. 1982a).

Approaches to determining utilization (Holechek et al.

1982a) have involved;

- evaluating differences between grazed and ungrazed

plots,

- evaluating differences before and after grazing,

- measurements involving correlation and regression

of factors related to utilization,

- general observation and comparison with

predetermined standards of use and

- ocular estimate methods which involve comparing

the amount of herbage inside and outside cages.

The advantages of this approach include the speed and

the fact that it provides information on location and the

degree a range is used in a given time period. Its major

disadvantage is that it does not indicate when a forage

species was used and how often it is used. This technique

20

does not account for large scale losses of plant parts

from weathering and trampling by animals other than those

of interest (Cook and Stoddart, 1953). Further still, for

actively growing forage, and regrowth after defoliation

can make accurate determination of utilization difficult.

Cook and Stoddart (1953) indicated that when forage is

actively growing and/or being used by more than one

herbivore, any utilization technique has severe

limitations. Under these conditions other procedures

should be selected for determination of botanical

composition of diet.

2.7.2 Direct observationDirect observation of the grazing animal is a widely

used procedure in studies of botanical composition of a

herbivore diet. The major advantages of direct

observation include simplicity, minor equipment

requirements and ease of use (Holechek et al. 1982a). The

problem associated with this method include difficulties

in species identification and quantifying information from

direct observation obtained from bite-count and feeding

minutes approaches. When the feeding minute technique is

employed, time spent grazing each species is quantified

and assumed to be proportional to the importance of the

species in the diet (Bjugstad et al. 1970). The

bite-count procedure differs in that number of bites taken

from each species, rather than the length of grazing time,

is recorded. This method may not apply to wild ungulates

21

which are often difficult to locate and approach closely

enough for accurate observations. These problems do not

occur in studies if tame animals are used for study. It

may be difficult to differentiate between mere nibbling

and active grazing, and only one animal can be observed at

a time even with tame animals.

Results from direct observation studies of tame

animals have been shown to be consistent with data from

esophageal fistulated animals (Sanders et al. 1980).

Sanders et al. (1980) reported that direct observation was

not practical for use on large brush infested pasture with

rough terrain. Factors that may influence the accuracy

and precision of direct observation procedure include the

degree of training of the observer, complexity of the

plant community present, and/or phenological development

of individual plants.

Diet selection is a complex behaviuoral act that is

influenced by several factors (Krueger et al. 1974).

Physiological condition, degree of hunger, topography,

other animals present and past experience, all influence

which and how much of individual plant species are

consumed. Therefore the previously mentioned factors can

be greatly altered by using artificially reared and

maintained animals.

2.7.3 Stomach content analysisStomach content analysis method of determining

botanical composition of animal diets is a common

22

procedure among wildlife researchers (Chippendale 1962,

Chamrad and Box 1964) . Stomach content analysis provides

information on what species are being consumed and gives

an indication of relative proportions consumed. The main

disadvantage of this procedure is that it involves killing

of animals and therefore, is restricted primarily to wild

animals with large populations. Difficulties also arise

from the fact the complete or partial digestion of some

material may alter the original proportions in the diet

and also make plant identification fragments difficult.

Tabulation of food item numbers, tabulations of frequency

of food item occurrences, volumetric measurement, and

weight measurement are methods that have been used to

evaluate stomach contents. Chamrad and Box (1964)

described a method which appears to be superior to other

methods in speed, accuracy and precision. The

microhistological technique by Spark and Malechek (1968),

and microscope point technique by Heady and Van Dyne

(1965) can be used to evaluate species composition by

weight.

A modification of this procedure has been reported by

Wilson et al. (1977) to avoid the problem of animal

sacrifice when stomach analysis is used to sample large

ruminant's diet. Tranquilization is used to immobilize

animals and rumen samples are taken with a trochar. The

resulting wound is sewn shut. However, layering of rumen

contents, effective tranquilization of animals and

infection by bacteria are problems associated with trochar

23

sampling. Further, due to the danger of death from

parasites, diseases and overdosing occurring often, the

technique may not be liberally applied to rare or

endangered species (Holechek et al. 1982a).

2.7.4 Microhistological techniqueMicrohistological technique has received greater use

for evaluating range herbivore food habits. The procedure

has several unique advantages that account for its

popularity in research (Croker 1959, Scother 1979) . These

include;

- non interference with the normal habits of animals,

- it permits practically unlimited sampling,

- there is no restriction on animal movement,

- it is of great value where animals graze over mixed

plant communities and

- actual sampling requires very little equipment.

Important disadvantages noted by Slater and Jones

(1971), Scother (1979), Vavra and Holechek (1980), Sanders

et al. (1980) include;

- accuracy may be a problem because forage species

passed in faeces are often not proportional to

those consumed,

- considerable equipment and labour are required for

actual analysis,

- an extensive reference plant collection is

required,

- an observer must have considerable experience or

24

training in order to accurately identify plant

fragments,

- some plant species are difficult to separate at the

species level and sometimes even at the genus

level,

- plant identification is tedious and time consuming,

- procedures of sample collection may bias the

results,

- some species of plants may become unidentifiable in

faeces and

- identification can be complicated by aging of

faecal material before sample collection.

Major points of interest concerning microhistological

techniques have been the influence of digestion, frequency

of observation and degree of training of observer on the

accuracy of diet composition determination. Hansen (1971)

reported good agreement between composition of ingested

and faecal material. Todd and Hansen (1973) indicated

that the relative number of plant fragments of each kind

of plant, remains similar in passing through the digestive

process. They suggest that digestion reduce the mean

weight of fragments rather than eliminating the whole

fragment. Sanders et al. (1980) compared the

microhistological technique and bite count method for

range animal diets. The experiment indicated that the two

methods gave similar results for estimating major

components of cattle diets. It was further noted that the

25

bite-count method could not be used on large bush-infested

areas with rough terrain whereas the other method could be

used under such conditions. Holechek and Gross (1982)

used seventeen forage species to determine the effects of

stage of maturity and skill of observation on the accuracy

of microhistological analysis for species composition of

hand compounded samples. Results indicated that growth

stage had little effect on the accuracy of estimation.

Five observers with specialized training and experience

obtained similarity indices for estimates of diets

composition of 95%, 88%, 85%, 93% and 75%. Poor accuracy

was recorded for the observer without specialized

training. The effect of observer was the most important

of the factors examined. Observers with specialized

training most accurately evaluated botanical composition

of each diet. In an experiment to study mastication

effects on cattle diets determined by microhistological

analysis, Bryan et al. (1983) reported that mastication

had no overall effect on diet composition. None of the

individual species was affected by mastication,

considerable variation occurred between observers in this

study. In general, the experiments pointed out that

fragments of plants survive the chewing and the digestive

process and can be identified and quantified in herbivore

faeces by microhistological technique. This method thus

remains the one of choice in the study of dietary

composition of free-ranging animals in mixed plant

communities.

26

2.7.5 Fistula techniqueEsophageal and rumen fistula techniques have

considerable advantages over some of the above methods in

that fistula enable the investigator to obtain naturally-

grazed samples. Both fistula are popular in research but

esophageal fistula is generally preferred to the rumen

fistula because rumen evacuation subjects animals to

abnormal conditions, is mainly limited to large animals

and is more laborious (Holechek et al. 1982a). Problems

associated with the use of the esophageal fistula include

contamination by rumen contents, incomplete recoveries,

high cost, and low sampling precision for individual

species in the diet.

27

CHAPTER THREESTUDY AREA

3.1 Location and physiographyGame Ranching Ltd (GRL), Athi River, is a privately

owned, mixed game and cattle ranch. GRL was initiated to

demonstrate the economic and environmental viability as

well as the social acceptability of game ranching. The

ranch occupies an area of 8,100 Ha. and is located 40 Km

South - East of Nairobi on the Athi Kapiti plains (Figure

1). Elevation varies between 1600 and 1700 M above sea-

level, latitude 0.1 30"S, and longitude 37 02"E (Stelfox

1985, MacDowell et al. 1988). The ranch is only 5 Km to

the North of Kajiado District, which is mainly used for

pastoralism and is bordered by unfenced private ranches.

It is separated from Nairobi National Park by Portland

Cement Ranch and the Nairobi - Namanga road.

Prior to 1981, GRL was operated as a cattle and

sheep ranch. Following the findings of research by

Hopcraft (1975) on productivity comparison between

Thomson Gazelle and cattle, and their relation to the

ecosystem, GRL applied to the Kenya Government for a

permit to operate it as a game ranch. To meet the

government regulations GRL had to complete several

modifications including a 50 Km chain-link fence, 2.6 M

in height along the perimeter to ensure the existence of

a closed system. This closed system was used in this

study to investigate the diets and habitat preferences of

cattle, and kongoni wildebeest.

28

38C 4(f 42°

KEY

□ Themeda grassland

Balanites tree grasland

E3 A. drepanolobium tree grassland

□ Balantes - Acacia tree grassland

■ Acacia woodland

) A. xanthophloea bushland ̂ ^

Figure 1: Geographical location of Game Ranching Ltd. and its vegetation types.

29

GRL falls within the eco-climatic zone four (IV),

the semi-arid zone, according to Pratt et al. (1966)

classification of the East African rangelands. The

following six vegetation types as shown in Figure 1 occur

in the ranch (MacDowell et al. 1988);

1. Themeda grassland

2. Balanites tree grassland

3. Balanites - Acacia tree grassland

4. Acacia drepanolobium dwarf tree grassland

5. Acacia woodland

6. Acacia xanthophloea bushland

The different habitat types did not have distinct

boundaries, however, the vegetation differences were

apparent. Themeda grassland areas are restricted only to

the ridge tops. Balanites tree grassland occurred on the

slopes. A. drepanolobium dwarf tree grassland, Balanites

- Acacia tree grassland and Acacia woodland occurred on

the lower slopes and areas with depressed topography. A.

xanthophloea bushland are associations restricted to a

seasonal stream bed at the northern part of the ranch.

Themeda grassland was characterized by absence of woody

plants. The Balanites tree grasslands were characterized

by widely spaced trees.

3.2 ClimateRainfall is bimodal and exhibits considerable

seasonal as well as year-to-year variation as shown in

Table 1. The long rainy season falls between March and

30

May, followed by a cool, cloudy and dry season from June

to September. The short rainy season extends from

October to December and is followed by a hot and sunny

dry period, which continues to the middle of March.

Average annual rainfall for 12 years starting in 1981 was

502 mm. During the eight month study period, January to

August the short rains extended to February as shown in

Figure 2. It should be noted however that despite the

fact that no rains were recorded during the month of

March, there were a lot of rains in some parts of the

ranch through to early April when rains actually stopped.

The usual long rains failed and as such January to April

and May to August were considered wet and dry seasons

respectively in this study. Due to the elevation,

temperature is characterized by warm days and cool

nights, maximum 24.9°C and minimum 13.7°C (MacDowell et

al. 1988).

Table 1: Mean and annual rainfall (mm) for the period

1981 to 1992.

Year 1981 1982 1983 1984 1985 1986 1987

Rainfall 422 473 435 349 698 468 345

Year 1988 1989 1990 1991 1992 mean

Rainfall 702 627 687 449 372 502

31

Rainfall (mm

)300

250

200

150

t_ i

i i «u- »

Mean (1981-1992)

Study period (1993)

Figure 2: Monthly rainfall during the study period compared to the mean monthly

rainfall for 12 years (1981 - 1992).

32

3.3 AnimalsThe wild herbivores of GRL include; kongoni,

Thomson's gazelle, grants' gazelle, Maasai giraffe,

impala, fringe-eared oryx, wildebeest and burchells

zebra. Results of a census conducted in the ranch during

the month of August by the ranch management are shown in

Table 2 below.

Table 2: Wild herbivore numbers and density per Km2 in the

study area.

Animal species Total Percentage Density/Km2

Wildebeest 949 34.3 11.7

Kongoni 583 21.1 7.2

Grants' gazelle 392 14.2 4.8

Thomson's gazelle 324 11.7 4.0

Burchells zebra 128 4.6 1.6

Fringe-eared oryx 109 3.9 1.3

Impala 96 3.5 1.2

Maasai giraffe 85 3.1 1.0

Ostrich 75 2.7 0.9

Eland 25 0.9 0.3

TOTAL* 2766 100.0 34.1

- Livestock were excluded from the census, Dairy herd 250, Penta herd 2360 (April-September) and Sheep 250.

33

CHAPTER FOURMATERIALS AND METHODS

4.1 Vegetation inventory and animal diets4.1.1 Herb layer biomass sampling

During the eight month study period, the sampling was

done six times, once every month except during the months

of March and July. Herb layer standing biomass was

obtained by clipping thirty 0.5 M 2 circular quadrats along

300 M transects in each vegetation type. Quadrats were

placed after every 10 M along the transects. This process

was repeated each time clipping was done, but on a

transect parallel and 4 M away from the previous one. The

herb layer was clipped to the ground level and separated

into species. Forbs were grouped together. The clipped

samples of each plant species were weighed in the field

and packed in paper bags. They were later dried to a

constant weight at 60°C and weighed. Dry weights were

determined for each habitat type and converted to

Kilogrammes per hectare.

4.1.2 Woody plant density and canopy cover samplingWoody plant density was estimated using the Point

Centred Quarter Method (PCQ) (Dieter and Heinz 1974) in

five habitat types. In each habitat two transects were

sampled. Themeda grassland was not sampled because it was

devoid of woody plants. A transect perpendicular to the

contours was established in each plot. A point 15 M along

34

each transect was selected for placing the PCQ stick.

Four sampling quarters was established by making two

lines, one parallel to the transect and the other

perpendicular to the transect at the PCQ stick. In each

quarter the distance between the nearest woody plant and

the stick was measured. Each plant was sampled for stem

diameter, and two perpendicular crown diameters. Canopy

cover was determined by calculating the means of the crown

diameters of each tree and hence their area. The crown

cover area of all the trees were totaled to give the total

cover and hence percentage cover calculated. The mean

area occupied by each tree was determined by using the

mean distance from the PCQ stick. Total area was then

calculated by multiplying the mean area by 120, the number

of quadrats sampled (4 x 30) in each vegetation type. The

density of the trees per hectare was then determined.

4.2 Dietary sampling4.2.1 Preparation of plant reference slides

In this study microhistological technique was used in

the analysis of dietary samples from the study animals.

Reference slides were prepared from plant species

collected from the study area. The plants were identified

in the field, collected, oven dried and finely ground

using a Willey Mill with 1 mm sieve. Slides were made

from the materials following the procedures described by

Cavender and Hansen (1970). A 5-10 gm sub-sample of the

reference material was soaked in a bleaching agent for

35

about 15 minutes. Soaking helped bleach the plant

pigments and softened the tissues during slide

preparation. Following this treatment the bleached

samples were then washed with running tap water over a 212

fim sieve for about 3 minutes to remove smaller plant

fragments, bleaching agent and any dirt. A small amount

of the material was then put near the end of a glass

microscope slide using a spatula and a template. Hoyer's

solution (Cavender and Hansen 1970) was added in small

quantities so as to cover the sample. A teasing needle

was used to mix the sample material with the solution.

The mixture was then spread evenly over an area large

enough to be covered by a 22 x 40 mm glass cover slip.

The cover slip was affixed and the slide was slowly heated

over an alcohol burner till the solution started to

bubble. The bottom part of the slide was immediately

pressed onto a wet cloth material to drive out air

bubbles. The cover slip was then gently, sealed to the

slide using a few drops of Hoyer's solution. Five slides

were prepared for each sample. The prepared slides were then placed in a rack and dried at 6 0°C for 4 8 hours and

stored.

4.2.2 Preparation of slides from faecal materialFor each of the animal species and for each month,

faecal samples were collected for 3 days each week during

a three week collection period from randomly selected

fresh dung/faeces in the field. Faecal samples for each

36

animal species were compounded from the weekly samples to

form one composite sample. Each composite sample was then

crushed, washed over a 600 /zm and 212 /xm sieves with

running tap water. The sample retained by the 212 fim

sieve was then dried and stored in sealed plastic bags.

For each composite sample five slides were made using the

procedures described above.

4.2.3 Slide quantificationFaecal slides were quantified as described by Hansen

et al. (1984) and Foppe (1984). Twenty identifiable

fields were used in the quantification. A lOOx

magnification was used on a movable stage microscope. The

recorded identifiable plant fragments were converted to

percent relative density, using the Hewlet-Packard

frequency/density tables, assuming that the relationship

between relative density and dry weight is highly

correlated (Sparks and Malechek 1968, Hansen et al. 1984,

Foppe 1984) . The percent relative density gives an

indication of the relative amount of different plant

species consumed (Hansen et al. 1984) . The frequency of

different plant species in each of the faecal slides was

determined and Relative Density (R.D) was calculated using

the formula (Hansen et al. 1984) below:

density of discerned fragments for a speciesR.D = ------------------------------------------------- x 100

densities of discerned fragments of all species

Tissue characteristics that were used in the

37

identification process included the characteristics of

epidermal cells, cell walls, stomatal shape, size and

arrangement of silica cells, druses, silica suberose,

trichomes and microhairs.

4.3 Animal density and habitat preferencesAnimal censuses were conducted once every month

during the study period (January to August, 1993) . Table

3 shows the areas covered by each habitat and their

respective areas that were sampled during the animal

censuses. In the Themeda grassland, Acacia woodland and

A. xanthophloea bushland habitat types which occupied less

than 10% of the Ranch each, one block was identified in

each, where total animal counts were done. In the other

habitat types that occupied more than 10%, each, of the

ranch (Balanites - Acacia tree, Balanites and A.

drepanolobi urn dwarf tree grasslands) two blocks were used

as shown in Figure 3. Each month, the censuses were

conducted for three consecutive days. The fieldwork was

done twice each day (in the morning and in the evening).

The morning census started at 6.30 am and ended at 9.30

am, while in the evening it started at 3.30 pm and ended

at 6.30 pm. The data recorded at the sites included

animal species, number, block (habitat type), date and

time of observation. Due to absence of clear boundaries

between the different habitat types, herds whose group

members appeared in transitional areas were assumed to

have been utilizing the block, hence counted. The areas

38

covered during the animal censuses were calculated by

using a map overlaid with grids and showing all the

locations of each vegetation type in the Ranch. The grids

on the map demarcated an area of 2.5 Km2.

Table 3: Total area in hectares covered by each habitat,

and their respective areas that were sampled

during the animal census.

HabitatTotal area Area sampled

(Ha) (%) (Ha) (%)

Balacc3 3937.50 48.61 275.00 6.98

Baltre2 2187.50 27.01 337.50 15.43

Accdrf4 1100.00 13.58 200.00 18.18

Accwdl5 500.00 6.17 112.50 22.50

Accxan6 225.00 2.78 75.00 33.33

Thtgsl1 150.00 1.85 75.00 50.00

Total 8100.00 975.00

1Themeda grassland, 2Balanites tree grassland, balanites - Acacia tree grassland, 4A. drepanolobium dwarf tree grassland, 5Acacia woodland and 6A. xanthophloea bushland.

4.4 Data analysis4.4.1 Seasonal diet preference

For each animal species and for each sampling period,

the dietary components at species level was tabulated

39

Figure 3: Map showing the areas sampled in each vegetation type for animals.

40

together with relative density and any changes in density

due to season. This was compared with plant species

available in the range so as to determine animal diet

preference by calculating their preference indices. Diet

preference indices were then calculated by using the

following equation adapted from Viljoen (1989).

Percentage in animal dietDiet preference index = ---------------------------

Percentage availability

4.4.2 Dietary overlapsDietary overlaps between animal species were

calculated using percent similarity index (S.I) (Oosting,

1956) .

E 2WiS.I. =■ i= 1

Y (ai+-ki)i =1

*100Where; Wi = Lesser percentage of

food category in the two diets being compared and

ai + bi = Sum of the percentages of food category in the two diets.

Comparisons were made between animals within seasons using

the above formula. The similarity index represents the

percentage of the diets that is identical or the

percentage of diets that is shared by two herbivores. It

shows the potential for competition between two

herbivores. Snedecor and Cochran (1967) and Hansen et al.

(1984) calculated a coefficient, Rs, as shown in the equation below from the relative densities of the plant

species common in the diets of two animals being compared

(Pair-wise comparisons).

41

Where;1-6 E rf2

i?s=- i= 1n (J23—1)

d = differences between pairedranks of common plant species,

n = number of pairs of observation and Rs = Spearman's rank correlation

coefficient.

Comparisons were made between animals within seasons. A

high Rs value (close to 1) indicates a strong correlation

for the order in which different plant species are

selected by the animals. A strong negative Rs value shows

a low potential for food competition by the animals being

compared. Test of significance for a positive Rs value

was by way of comparison with table values at n-2 degrees

of freedom (Snedecor and Cochran 1967) . Using the above

two methods, animal diets were compared two at a time for

each season. Significant positive values were interpreted

to mean strong correlations for the order in which any two

herbivores select the same plant species that constitute

their diets. Trophic diversity indices (H") were

calculated using the formula below to indicate food niche

breadth using Shannon's diversity index (1948).

H //=~y^ NilnNi where Ni is the proportion of the ithn=i food item in the diet.

Trophic diversity index indicates variety and evenness of

components in the diet. The index decreases as an animal

relies on fewer food categories for most of their diets

(Hurtubia 1973). High trophic diversity indices indicate

that an animal does not rely of fewer major food

categories for most of their diets, but is able to feed on

many categories. Therefore, the herbivore with high

trophic diversities is potentially better able to adopt

42

its diets to changes in plant composition.

4.4.3 Animal density and habitat preferencesAnimal density per Km2 was calculated as shown below:

Animals counted in each blockAnimal density (N/Km2) = ---------------------------- x 100

Total block area (Ha)

The total number of each species was obtained by

multiplying its density in each vegetation type by the

total area occupied by the vegetation type and summing up

the totals of each vegetation type. The distribution of

kongoni and wildebeest in all the habitat types were

determined by monthly ground sample counts. To evaluate

the animals' habitat preference, the relative densities of

each animal species were calculated on seasonal basis. A

Preference Index (P.I) was calculated for each species in

each habitat type using an equation adopted from Hillman

(1979) as shown below;

1 . If, nx > ax Nt At'

then, P. I (x) 1__y { nx_ ax)ax Nt At At

2. If, then,p.l(x) =•Nt At ax Nt

At

axAt

Where; P.I = Habitat Preference Index,nx = the number of animal species expected in

habitat "X",Nt = the total number of each animal species

expected in the ranch, ax = the surface area of habitat "X" (Km2),At = area (Km2) covered during the census in

habitat "X',nx/Nt = proportion of animal species expected in

habitat "X" andax/At = the proportion of the total area covered by

habitat "X".

43

The values obtained range from -1.0 to +1.0, indicating

the least and the most preferred habitats respectively. A

value of zero indicates a random association and a value

of -1.0 indicates that the habitat was not utilized at

all.

The data was analyzed using a split-split plot

design. Treatment means were separated using Duncan's

Multiple Range Test and statistical analysis were done

using SAS procedures (SAS Inc.,1987) and graphics enhanced

using Freelance Program (Lotus Development Corporation

1991).

44

CHAPTER FIVERESULTS AND DISCUSSION

5.1 Vegetation and animal diets5.1.1 Woody plant cover and density

The results of the ten PCQ transects showed clear

differences in canopy cover and tree density between

transects representing the different vegetation types

(habitats). In the Balanites tree grassland the canopy

cover was below 2% (Table 4). In this vegetation type

Table 4: Mean absolute and relative canopy cover (%) and

density (number of plants per hectare) for

Balanites tree grassland.

Dominant speciesDensity ------ „----------- 1Cover

Absolute Relative Absolute Relative

B. glabra 13.67 92.92 1.37 97.89A. drepanolobium 0.76 5.42 0.02 1.31A. xanthophloea 0.21 0.72 0.01 0.78A. Stuhlmanii 0.07 0.42 0.01 0.02All species 14.70 100.00 1.40 100.00

Balanites glabra contributed 97.89% of the woody plant

cover while the other species contributed only 2.11%.

This vegetation type had a lower density of trees (14.7

trees per hectare) with B. glabra contributing the highest percentage 92.92%.

45

As shown in Table 5 the Balanites - Acacia tree

grassland canopy cover was dominated by A. drepanolobium

contributing 69.17% while A. stuhlmanii and B. glabra

each contributed 15.14% and 15.47%, respectively. This vegetation type had the highest tree density of 603.98

trees per hectare, with A. drepanolobium contributing the

highest density (86.67%).

Table 5: Mean absolute and relative canopy cover (%) and

density (number of plants per hectare) for

Balanites - Acacia tree grassland.

Dominant speciesDensity Cover

Absolute Relative Absolute Absolute

A . drepanolobi urn 523.44 86.67 7.62 69.17B. glabra 60.41 10.00 1.73 15.47A. stuhlmanii 17.62 2.92 1.71 15.14Commiphora spp 2.51 0.42 0.02 0.23All species 603.98 100.00 11.07 100.00

The A. drepanolobium dwarf tree grassland was

dominated by A . drepanolobium which contributed 94.15%

(Table 6) and 91.25% cover and trees per hectare,

respectively, while A. stuhlmanii contributed 5.79% cover

and 7.92% trees per hectare. The Acacia woodland had the

highest number of dominant trees, with a total cover of

12.63% and high tree density of 456.41 trees per hectare

as shown in Table 7. A. drepanolobium contributed 47.09%

46

of the cover and 41.67% of tree density. The A.

xanthophloea bushland had the highest woody cover of

122.92% and tree density of 215.33 tree per hectare

(Table 8). Tall trees above 8 M occurred in this

vegetation type. This area occurred long a watercourse,

where the ground water was close to the surface. A.

xanthophloea contributed the highest percentage both in

cover and density, 95.40% and 72.92%, respectively.

Table 6: Mean absolute and relative canopy cover (%) and

density (number of plants per hectare) for A.

drepanolobium dwarf tree grassland.

Dominant speciesDensity Cover

Absolute Relative Absolute Relative

A . drepanolobi um 168.04 91.25 2.56 94.15

A. stuhlmanii 14.98 7.92 0.15 5.79

B. glabra 1.49 0.83 0.01 0.06

All species 184.52 100.00 2.72 100.00

Table 7: Mean absolute and relative canopy cover (%) and

density (number of plants per hectare) for

Acacia woodland.

Dominant speciesDensity Cover

Absolute Relative Absolute Relative

A. drepanolobium 187.49 41.67 5.93 47.09

A . xanthophloea 177.03 38.33 4.66 36.88

A. seyal 57.64 12.50 1.02 8.06

B. glabra 26.62 5.83 0.92 7.32

Commiphora spp 7.61 1.67 0.09 0.72

All species 456.41 100.00 12.63 100.00

48

Table 8: Mean absolute and relative canopy cover (%) and

density (number of plants per hectare) for A . xanthophloea bushland

Dominant speciesDensity Cover

Absolute Relative Absolute Relative

A. xanthophloea 157.04 72.92 117.07 95.40B. glabra 38.70 17.92 3.96 3.08Dombeya spp 9.78 4.58 0.88 0.73A . seyal 7.34 3.34 0.88 0.67

A. stuhlmanii 1.76 0.83 0.12 0.11Commiphora spp 0.88 0.42 0.01 0.01All species 215.33 100.00 122.92 100.00

Analysis of variance of the density and percent

canopy cover of the trees in all the habitats sampled,

revealed that both the density and canopy cover was

dependent upon the habitat (environment), tree species

and their interactions. Evidence for this is shown by a

significant differences in density between habitats

(p4,45=92 • 49 df=4 p<0.05) , trees (F6 45=237.06 df = 6 p<0.05)

and habitat - tree interaction (F24 4 5 = 93.81 df=24 p<0.05)

in the analysis as shown in Table 9. Similarly there

were significant differences in percent canopy cover

between habitats (F4 45=429.10 df=4 p<0.05), trees

(F6,45=439.24 df=6 p<0.05) and habitat - tree interaction

(F24,45 = 4 2 7.54 df = 24 p<0.05) (Table 10).

49

Table 9: Analysis of variance of density of trees per

hectare in the study area.

Source F SS MS F Value

CORRECTED TOTAL 69 687932.24

PLOT 1 29.35 29.35 0.17ns

HABITAT 4 62460.70 15615.17 92.49*

TREE 6 240128.41 40021.40 237.06*

PLOT*HABITAT 4 351.17 87.79 0.52nsPLOT*TREE 6 825.13 137.52 0.81nsHABITAT*TREE 24 380085.66 15836.90 93.81*

- significant (p<0. 05) , ns - non significant.

Table 10: Analysis of variance of percent canopy cover of

woody plants in the study area.

Source DF SS MS F Value

CORRECTED TOTAL 69 26472.04

PLOT 1 4.11 4.11 2.28nsHABITAT 4 3101.96 775.49 429.10*TREE 6 4762.94 793.82 439.24*PLOT*HABITAT 4 11.37 2.84 1.57nsPLOT*TREE 6 3.84 0.64 0.35nsHABITAT*TREE 24 18544.43 772.68 427.54*

significant (p<0.05), ns - non significant.

50

The mean percent canopy cover varied greatly from

habitat to habitat and from tree to tree. It ranged from

0.20 to 17.56 for Balanites tree grassland and A .

xanthophloea bushland habitats respectively (Table 11),

and from 0.02 to 24.35 for Commiphora spp and A.

xanthophloea tree species respectively (Table 12). The

tree density per hectare likewise varied greatly from

habitat to habitat. It ranged from 2.10 in the B. glabra

tree grassland to 86.28 in Balanites - Acacia tree

grassland (Table 11). The same phenomenon likewise

varied from tree to tree. It ranged from 1.96 for

Dombeya spp. to 175.95 for A. drepanolobium (Table 12).

Table 11: Mean density per hectare and percent canopy

cover of trees in each habitat in the study

area.

Habitat Density Canopy Cover n

Baltre2 2.10Id* 0.200c* 14Balacc3 86.284a 1.581b 14Accdrf4 26.359c 0 .388c 14

Accwdl5 65.200b 1.804b 14

Accxan6 30.786c 17.560a 14

means followed by the same letter in the same column are not significantly different at P<0.05. 1Themeda grassland, 2Balanites tree grassland, 3Balanites - Acacia tree grassland, 4 A. drepanolobium dwarf tree grassland, 5Acacia woodland and 6A. xanthophloea bushland.

51

Table 12: Mean density per hectare and percent canopy

cover of each tree species in the study area.

Species Density Cover n

A . drepanolobium 175.949a* 3.226b* 10A . seyal 12.996d 0.379cd 10A. xanthophloea 66.856b 24.348a 10

B. glabra 28.178c 1.596c 10Commiphora spp 2.200d 0.024d 10A. stuhlmanii 6.886d 0.3 96cd 10

Dombeya spp. 1.956d 0.176d 10

means followed by the same letter in the same column are not significantly different at P<0.05.

5.1.2 Herb layerThe herbaceous standing biomass remained high

during the study period (Table 13). All the habitats,

except Themeda grassland, had very high standing biomass

both during the wet and dry seasons, but were generally

lower during the dry season. Themeda grassland had the

lowest mean standing biomass of 789 and 495 Kg/ha. during

the wet and dry seasons, respectively. Balanites tree

grassland, which was heavily utilized by the study

animals, had 1838 and 1386 Kg/ha. mean standing biomass

during the wet and dry seasons, respectively. Other

habitats had higher mean standing biomass than Balanites

tree grassland irrespective of the season, with Acacia

woodland always having the highest. The change in

percentage standing biomass of dominant grass species

52

varied with habitat type. Grasses with more than 10%

standing biomass were termed dominant, while those with

between 1% and 10% were termed subdominant.

Table 13: Mean herbaceous standing biomass in Kilogrammes

per hectare for each habitat during the wet and

dry seasons.

Habitat Thtgsl1 Baltre2 Balacc3 Accdrf4 Accwdl5 Accxan6

Wet 789 1838 2005 2456 3130 2585Dry 495 1386 1836 2218 2458 1768

1Themeda grassland, 2Balanites tree grassland, 3Balanites - Acacia tree grassland, 4A. drepanolobium dwarf tree grassland, 5Acacia woodland and 6A. xanthophloea bushland.

Themeda triandra, Digitaria macroblephara and

Penisetum mezianum were the three most dominant species

in the Themeda grassland, Balanites tree grassland,

Balanites - Acacia tree grassland and in Acacia woodland

vegetation types (tables 14, 15, 16 and 18)'. Ischaemum

afrum, Lintonia nutans, and P. mezianum were the

dominant species in the A. drepanolobium dwarf tree

grassland (Table 17). In the A. xanthophloea bushland

Penisetum stramineum, P. mezianum, D. macroblephara and

P. maximum were the dominant species, (Table 19) . The

subdominant species, unlike the dominant ones, varied

from one vegetation type to another and showed no general

pattern. They ranged from D. macroblephara in the A.

53

drepanolobium dwarf tree grassland in the depressed areas

to Cynodon dactylon in Themeda grassland at the ridge

tops (Tables 14 - 19) .

Table 14: Seasonal percent standing biomass of dominant

grasses (>1%) and forbs in the Themeda

grassland.

SEASON

Plant Species Wet Dry

Mean SD Mean SD

D. macroblephara 29 .90 5 .18 22 00 5 70

T. triandra 11. 40 2 .46 12 10 1 98

P. mezianum 10. 60 8 .24 17 80 17 60

C. dactylon 9. 90 8 . 74 10 60 5 17

Harpachne schimperi 7. 13 1 .69 5 20 3 67

Panicum spp 6 .33 5 . 98 5 26 4 64

Bothriochloa inscupta 5.53 2 . 01 4 44 1 83

P. stramineum 2. 50 3 .12 9 74 4 38

Microchloa kunthii 1.26 1 .64 0 00 0 00

Sedges 1 01 1 . 74 0 00 0 00

Chi oris gayana 0 09 0 .16 1 15 2 00

Penisetum masaicum 0. 00 0 .00 2 .92 5 06

Forbs 12 .70 2 .62 7 .43 5 .73

54

Table 15: Seasonal percent standing biomass of dominant

grasses (>1%) and forbs in the Balanites tree

grassland.

Plant Species

SEASON

Wet Dry

Mean SD Mean SD

T. triandra 42.40 4.77 36.10 1.88

D . macroblephara 21.00 4.35 18.40 1.64

P. mezianum 14.50 5.91 16.90 8.97

B. inscupta 3.69 1.09 2.79 1.08

Hyperrhenia spp 3.66 3.82 16.40 4.58

P. stramineum 1.61 2.80 0.00 0.00

Aristida spp 1.61 1.23 1.48 0.50

H. schimperi 1.39 1.32 0.16 0.29

I. afrum 1.00 0.88 0.00 0.00

C. dactylon 0.88 0.78 1.75 0.95

Forbs 5.26 1.78 5.30 3.67

55

Table 16: Seasonal percent standing biomass of dominant

grassland (>1%) and forbs in the Balanites -

Acacia tree grassland.

Plant Species

SEASON

Wet Dry

Mean SD Mean SD

P. mezianum 26.80 7.49 36.80 5.44

T. triandra 25.10 4.42 21.20 0.19

D. macroblephara 15.40 4.33 17.10 1.48

P. masaicum 9.17 11.30 0.00 0.00

P. stramineum 7.60 7.53 8.18 2.43

B. inscupta 2.99 2.61 0.00 0.00

Hyperrhenia spp 2.15 3.73 0.00 0.00

C . gayana 1.83 2.95 0.00 0.00

L. nutans 1.66 2.69 8.05 3.40

C. dactylon 1.61 2.17 1.14 1.09

I. afrum 0.44 0.77 5.77 1.45

Forbs 3.83 1.19 1.55 0.77

56

Table 17: Seasonal percent standing biomass of dominant

grasses (>1%) and forbs in the A. drepanolobium

dwarf tree grassland.

Plant Species

SEASON

Wet Dry

Mean SD Mean SD

D. macroblephara 15.60 15.80 3.42 1.41

P. mezianum 14.50 13.90 7.40 1.76

P. stramineum 13.80 7.67 11.70 6.26

L. nutans 12.50 12.20 14.70 0.25

I. afrum 11.40 8.15 30.00 2.68

P. masaicum 2.65 3.34 0.00 0.00

B. inscupta 1.22 1.82 0.00 0.00

Sporobolus pellucidus 1.07 1.29 0.00 0.00

C. dactylon 0.45 0.52 7.12 6.18

Forbs 25.10 9.61 24.70 3.36

57

Table 18: Seasonal percent standing biomass of dominant

grasses (>1%) and forbs in the Acacia woodland.

Plant Species

SEASON

Wet Dry

Mean SD Mean SD

D. macroblephara 20.80 1.97 29.70 3.12

P. mezianum 19.30 2.11 21.30 3.24

T. triandra 17.00 2.28 14.90 5.80

I . afrum 14.40 3.65 9.95 5.72

L. nutans 7.82 7.12 12.30 6.17

P. stramineum 4.99 4.45 4.31 3.80

Forbs 14.70 5.00 6.42 5.08

58

Table 19: Seasonal percent standing biomass of dominant

grasses (>1%) and forbs in the A. xanthophloea

bushland.

Plant Species

SEASON

Wet Dry

Mean SD Mean SD

P. stramineum 29.40 7.88 24.00 4.47

P. mezianum 19.30 6.15 22.40 2.60

D. macroblephara 18.80 4.63 20.60 6.91

P. maximum 10.30 1.07 16.20 3.05

L. nutans 4.48 5.27 7.40 5.43

I. afrum 3.39 3.16 1.90 1.70

C. dactylon 0.75 0.33 1.90 0.90

C. gayana 0.00 0.00 1.07 0.93

Forbs 12.20 5.82 4.24 0.32

Table 20 shows the analysis of variance of seasonal

herbaceous mean percent standing biomass above 1%. There

was no significant variation in percent standing biomass

among plant species between the wet and the dry season.

However the percent standing biomass varied from habitat

to habitat (F5 62=7.04 df=5 p<0.05) . It was highest in

Acacia woodland and lowest in Themeda grassland (Table

21) . The standing biomass varied greatly among the plant

species (F1862=24.25 df = 18 p<0.05). It was highest

among three dominant grasses; P. mezianum, D.

59

macroblephara and T. triandra, where it ranged from 19%

to 22.6% respectively. The standing biomass of forbs in

general was lower (10.31%) than grasses, but was higher

than for some individual grasses as shown in Table 22.

The percent standing biomass of herbaceous plants is

dependent on the interaction between the plant and the

environment. Evidence for this is shown by a significant

interaction term between habitat and herb in the analysis

as shown in Table 20 (F38 61=7.00 df=38 p<0.05).

Table 20: Analysis of variance of seasonal percent

standing biomass of dominant grasses (>1%) and

forbs.

Source DF SS MS F Value

CORRECTED TOTAL 123 11067.18

SEASON 1 0.22 0.23 0.02ns

HABITAT 5 487.60 97.52 7.04*

PLANT 18 6047.14 335.95 24.25*

HABITAT*PLANT 38 3687.20 97.03 7.00*

- significant (p<0.05) , ns - non significant.

60

Table 21: Mean percent standing biomass of dominant

grasses (>1%) and forbs in each habitat.

Habitat Mean standing biomass n

Accwdl5 14.171a* 14

Accxan6 11.052b 18

Accdrf4 9.892bc 20

Baltre2 8.943bc 22

Balacc3 8.282c 24

Thtgsl1 7.598c 26

* Means followed by the same letter are not significantly different (p<0.05). 1Themeda grassland, 2Balanites tree grassland, 3Balanites - Acacia tree grassland,4A. drepanolobium dwarf tree grassland, 5Acacia woodland and eA. xanthophloea bushland.

61

Table 22: Mean seasonal percent standing biomass of

dominant grasses (>1%) and forbs.

Plant species Mean n

T. triandra 22.582a* 8

D . macroblephara 19.444a 12

P. mezianum 19.026a 12

P . maximum 13.310b 2

Forbs 10.308b 12

P. stramineum 9.84 Obc 12

L. nutans 8.634bcd 8

I. afrum 7.845bcde 10

Panicum spp 5.800bcde 2

Hyperrhenia spp 5.575cdef 4

C. dactylon 3.618def 10

H. schimperi 3.475def 4

B. inscupta 2.586ef 8

P. masaicum 2.460ef 6

Aristida spp 1.545f 2

C . gayana 0.693f 6

M. kunthii 0.635f 2

Sporobolus pellucidus 0.535f 2

Sedges 0.505f 2

Means followed by the same letter are not significantlydifferent (p<0.05).

62

5.1.3 Diet composition by plant speciesThe botanical composition in the animals' diets

during the wet and dry seasons are shown in Tables 23 and

24. During the wet season, the three most prominent

grass species in each animal species diet (Table 23 and

Figure 4 (A)) and their means were as follows:

Cattle: T. triandra, D. macroblephara and P.

mezianum, each contributed 29.5%, 24.7% and

15.9% respectively. Their combined

contribution was 70.2%. Other monocot

species combined contributed 25.8% while

browse contribution was 4.0%.

Kongoni: D. macroblephara, T. triandra and P.

mezianum were the prominent grass species.

Each species made up 20.8%, 21.1% and 19.4%

respectively. Their combined contribution to

the diet was 61.3%. Other grass species'

combined diet contribution was 37.4%, whereas

browse contributed only 1.2%.

Wildebeest: T. triandra, D. macroblephara and P. mezianum

made up 23.7%, 21.1%, and 18.7% of the diet

respectively. Their combined contribution

being 63.4%. Other grass species contributed

34.7% while browse component of the diet was

only 1.9%.

63

Table 23: Mean percent relative densities of the botanical composition of the diets of the study animals during the wet season (January-April, 1993) ._____________________________________________________

SpeciesCATTLE KONGONI WILDEBEEST

Jan Feb Mar Apr Jan Feb Mar Apr Jan Feb Mar Apr

T. triandra 33.56 30.32 26.73 27.32 19.76 21.10 21.22 22.30 28.00 26.24 21.85 18.71D. macroblephara 25.79 27.59 27.44 18.17 26.13 18.21 16.37 22.60 20.38 15.74 20.62 27.48P. mezianum 13.00 14.39 15.03 21.31 19.57 23.61 16.80 17.73 21.65 17.82 21.24 13.92i3. stramineum 2.60 2.30 3.68 2.33 6.39 7.15 15.70 8.96 1.59 5.16 5.03 5.11I. afrum 1.64 2.08 8.05 6.58 6.51 4.46 5.44 1.25 3.68 7.40 10.47Other grasses* 2.09 2.62 2.20 2.78 3.42 6.69 3.68 4.01 7.52 10.82 3.01 3.43C. dactylon 4.25 4.01 4.45 6.29 2.66 2.74 3.43 3.49 3.07 2.97 3.25 1.97L. nutans 0.08 1.67 1.49 1.61 4.25 4.65 4.28 5.35 1.96 2.70 6.37 3.88Browse 2.80 2.39 5.20 5.76 • 1.18 1.78 1.90 2.93 0.98 0.86 2.83B. inscupta 3.34 2.17 1.58 0.70 2.35 3.31 0.73 1.03 1.07 2.19 1.60 2.63Aristida spp 3.45 2.76 0.80 0.49 1.78 0.86 1.21 0.60 1.27 1.67 0.75 1.17C. gayana 1.65 0.74 0.93 0.78 1.85 1.01 0.69 0.18 3.09 0.85 2.90 0.60Brachiaria spp 2.42 1.39 2.26 0.59 1.81 0.70 2.08 0.32 0.56 1.37 0.44 0.88Hyperrhenia spp 0.67 0.46 0.25 0.41 • 0.19 0.98 1.15 1.72 0.49 0.90 4.51P. masaicum 0.34 0.33 0.90 0.25 1.02 2.55 3.05 0.00 1.47 0.46 1.27H. contortus 0.50 0.60 1.72 0.18 0.18 0.46 1.15 0.62 0.67 2.56 2.16 •H. schimperi 0.90 1.17 1.86 0.63 0.83 0.47 0.81 0.32 1.24 1.32 0.09 0.43C. ciliaris 1.14 0.60 0.95 0.69 1.34 0.42 1.48 0.52 0.89 0.51 0.37 •S. pellucides 1.19 1.59 0.46 0.95 0.08 0.74 0.36 0.22 • 0.22 . •Sedges . 0.74 . 0.71 . . 0.23 0.23 0.45 • 0.74 0.72Panicum spp 0.23 0.53 • • • • • • • •

* - unidentified grasses in the diet and . (dot) not observed in the diet.

64

Per

cent

in d

iet

Per

cent

in d

iet

30

20

10

0

Figure 4: Prominent grass species and browse in animal diets during the wet (A) and dry (B) seasons at Game Ranching Ltd., Athi River, Kenya. 1993.

The mean percent relative densities of the botanical

composition of animal diets during the dry season are

shown in Table 24. The three most dominant grass

species, other grasses and browse contribution to the

diet (Figure 4 (B)) and their means were as follows;

Cattle: T. triandra, D. macroblephara and P. mezianum

were prominent grass species. Their

contribution to diets were 22.3%, 19.4% and

18.7%, respectively. The three made up 60.3%

of the diet. Other grass and browse species

each contributed 31.6% and 8.1% to the diet.

Kongoni: D. macroblephara, T. triandra and P.

mezianum were the three most abundant plant

species in the diet. These grasses

contributed 22.2%, 16.4% and 16.1%,

respectively thus constituting 54.7% of the

diet. Other grasses contributed 41.9%,

whereas browse component was 3.8% of the

diet.

Wildebeest: P. mezianum, D. macroblephara and T.

triandra, each contributed 20.3%, 20.8% and

15.8% respectively to the diet, thus making

up 57.0% of the diet. Again browse

contributed only 3.8%, whereas other grasses

contributed 39.2% to the diet.

66

Table 24: Mean percent relative densities of the botanical composition of the diets of the study animalsduring the dry season (May - August, 1993).

SpeciesCATTLE KONGONI WILDEBEEST

May Jun Jul Aug May Jun Jul Aug May Jun Jul AugD. macroblephara 20.79 18.85 19.19 18.76 22.39 22.62 23.21 20.59 24.63 19.71 21.82 16.91P. mezianum 16.60 18.04 19.69 20.28 14.64 15.47 18.11 16.14 22.20 21.57 17.33 20.24T. triandra 24.17 22.11 21.15 21.71 15.83 15.37 18.19 16.32 14.57 16.00 16.38 16.45I. afrum 6.13 4.89 7.65 8.47 13.79 13.61 12.78 10.62 8.29 9.93 8.80 8.44L. nutans 4.41 7.70 5.57 7.73 9.90 10.25 7.28 9.35 6.30 6.18 5.88 6.62P. stramineum 7.33 8.72 4.36 3.88 4.08 9.30 6.08 8.04 3.19 6.93 8.59 7.32C. dactylon 3.94 2.23 3.79 3.29 8.60 5.11 5.91 6.33 6.30 8.11 6.41 12.06Browse 8.42 9.21 5.33 9.33 6.62 1.70 2.05 3.09 3.61 3.73 4.43 3.50Other grasses* 3.52 3.60 4.83 4.57 2.69 4.57 3.05 4.29 2.94 4.25 6.70 5.64Sedges 0.49 1.01 3.72 0.14 0.55 . 3.24 1.49 0.40 • 0.11Hyperrhenia spp 1.58 . 1.78 • . . 0.21 • 1.22 0.20 1.56 1.13Brachiaria spp 0.54 0.18 1.54 0.73 0.30 . 0.42 . 0.17 0.79 0.16 .C. gayana 0.63 0.55 0.19 0.72 0.08 0.36 0.42 0.26 0.45 0.49 0.63 •B. inscupta 0.26 0.45 0.29 . 0.14 0.19 1.37 0.27 0.45 0.80 . 0.09C. ciliaris . 0.47 0.92 0.55 . 0.09 0.10 • 0.35 0.45 0.17 0.72&S. pellucides 0.49 1.50 . . 0.23 . 0.21 . 0.86 0.34 . .Aristida spp 0.34 . . 0.15 . 0.31 . 1.98 • 0.10 0.10P. masaicum . . . . 0.16 • 0.21 1.37 0.47 • 0.35 0.36H. contortus . 0.48 . . 0.26 0.54 . • • • 0.70 0.29H. schimperi 0.36 • • • • 0.27 0.10 0.09 0.53 0.11 • •

* - unidentified grasses in the diet and . (dot) not observed in the diet.

67

y v

Table 25: Mean diet preference indices* of cattle, kongoni andwildebeest with percent plant availability (PA) during the wet and dry seasons, at GRL, Athi River 1993.

Plant species

SEASONWet Dry

Ct Kg Wb PA Ct Kg Wb PA

Brachiaria spp 10.72 6.30 12.38 0.30 # . # 0.07H. schimperi 5.68 2.93 4.56 0.43 2.99 2.50 5.35 0.06C. dactylon 5.68 3.39 2.71 1.18 1.36 2.91 4.10 2.35C. gayana 3.27 4.39 • 0.97 10.40 • . 0.04Aristida spp. 4.23 2.54 4.28 0.56 0.29 0.13 1.81 0.31Sedges 3.79 0.31 3.64 0.13 • • • 0.02D . macrohlephara 1.42 1.28 1.20 17.46 1.20 1.35 1.26 16.18I. afrum 0.65 2.10 1.21 3.64 0.73 1.41 1.00 8.99P. stramineum 0.85 2.28 0.80 6.67 0.94 1.05 0.87 6.88S. pellucides 5.18 1.37 0.20 0.31 . . . 0.07T. triandra 1.21 0.84 0.96 25.41 1.12 0.78 0.77 20.25B. inscupta 0.87 1.07 1.62 2.54 0.34 0.37 0.66 0.61P. mezianum 0.77 1.00 0.89 20.71 0.71 0.60 0.84 25.79Panicum spp. 1.14 0.00 3.08 0.18 • • • 0.06Browse 0.43 0.11 0.28 8.63 1.31 0.54 0.53 6.97L. nutans . . 3.61 0.89 1.30 0.84 7.85P. masaicum 0.19 0.55 0.83 4.79 . • • 0.02Hyperrhenia spp. • • • 1.98 0.15 0.00 0.29 3.47

*Preference index = Percent in diet/percent availability, Kg - kongoni, Wb - wildebeest, Ct - cattle and PA - plant availability.

The seasonal diet preference indices for cattle,

kongoni and wildebeest with percent plant availability

in the ranch are shown in Table 25. Analysis of variance

of mean diet preference indices of the plant species

eaten by the animals (Table 26) revealed that the diet

preference indices differed significantly between the

animals (F218=8.39, df=2, p<0.05). Brachiaria species

68

*

Table 26: Analysis of variance of mean diet preferences

of cattle, kongoni and wildebeest.

Source DF SS MS F Value

CORRECTED TOTAL 83 454 14

SEASON 1 20 68 20 68 33 72*

ANIMAL 2 10 29 5 14 8 39*

DIET 17 315 96 18 58 30 30

ANIMAL*DIET 33 53 .02 1 60 2 62*

SEASON*DIET 10 40 .88 4 08 6 66*

SEASON* ANIMAL 2 2 .24 1 12 1 83ns

- significant (p<0.05), ns - non significant.

recorded the highest diet preference index, followed by

C. gayana, while Hyperrhenia species had the lowest index

(Table 27). The diet preference indices for animals

differed from one season to the other (Flie=33.72, df = l,

p<0.05). It tended to be high during the wet season and

low during the dry season indicating that the degree of

selectivity was low during the dry season. Similarly,

the diet preference indices varied significantly from one

plant species to another (F17 18=30.30 , df=17, p<0.05).

The diet preferences seemed to be dependent on the

interactions between season and plant species and on the

interactions between animals and plant species and not on

animal - season interaction. This is supported by a

significant season - diet and animal - diet interaction

terms in our model (F1018=6.66 df = 10, p=0.05 and

69

F33,i8=2.62/ df=33, p<0.05 respectively), while animal -

season interaction was non significant.

Table 27: Mean diet preferences of cattle, kongoni and

wildebeest.

Diet Mean diet preference n

Brachiaria spp 9.800a* 3

C. gayana 6.020b 3

H. schimperi 4.002c 6

C. dactylon 3.358cd 6

Sedges 2.580cde 3

S. pellucidus 2.250cdef 3

Aristida spp 2.213cdef 6

Panicum spp 1 .407edf 3

D. macroblephara 1.285ef 6

I. afrum 1.183ef 6

P. stramineum 1.132ef 6

L. nutans l.OlOef 3

T. triandra 0.947ef 6

B. inscupta 0.822ef 6

P. mezianum 0 .802ef 6

Browse 0 .533ef 6

P. masaicum 0 .523ef 3

Hyperrhenia spp 0 .147ef 3

Means followed by the same letter are not significantly different at p<0.05.

70

5.1.4 Seasonal dietary overlaps between animalsDietary overlaps during the rainy season were all

above 70% and significant (P<0.05), as shown in Table 28

and illustrated in Figure 5. Cattle and kongoni diets

had overlaps ranging from 72.59% to 79.56% with a mean

overlap of 76.06%. Correlation coefficients were

significant (p<0.05) and ranged from +0.46 to +0.87.

Cattle and wildebeest diet overlaps ranged from 74.56% to

78.74%, with a mean overlap of 77.30%. All correlation

coefficients were significant (P<0.05) and ranged from

+0.62 to +0.72. Diet overlaps between kongoni and

wildebeest ranged from 77.24% to 82.06% with a mean

overlap of 80.83%. Correlation coefficients were

significant during all periods and ranged from +0.64 to

+0.90 (p<0.05).

71

correlation coefficients (Rs) between cattle

Table 28: Percent similarity indices (S.I) and

and wildlife diets during the wet season.

Animal combination Period S.I Rs n t

Cattle Vs kongoni Jan 76.14 0.46 20 2.218*

Feb 72.59 0.87 21 7.683*

Mar 75.99 0.78 20 5.253*

Apr 79.56 0.66 20 3.699*

Mean 76.06 ± 2.46

Cattle Vs wildebeest Jan 78.74 0.62 21 3.405*

Feb 78.30 0.72 21 4.507*

Mar 77.64 0.62 20 3.383*

Apr 74.56 0.68 20 3.940*

Mean 77.30 ± 1.63

Kongoni Vs wildebeest Jan 77.24 0.64 21 3.669*

Feb 82.97 0.80 20 5.726*

Mar 81.08 0.76 20 5.031*

Apr 82.06 0.90 20 8.890*

Mean 80.83 ± 2.18

* - significant (p<0.05).

72

76.1

Cattle Vs Kongonl wet season

Cattle Vs Kongonl dry season

Cattle Vs Wildebeest wet season

83 .

6.8

Cattle Vs Wildebeest dry season

80.8

19.2

85.8

14.2

Kongonl Vs Wildebeest wet season

Kongonl Vs Wildebeest dry season

Diet overlap No diet overlap

Figure 5: Percent animal diet overlaps during the wet ( January - April) and dry(May - August) seasons at Game Ranching Ltd. Athi River, Kenya. 1993.

73

Dietary overlaps exhibited during the dry season

were all above 75%, (Table 29 and Figure 5). Cattle and

Kongoni diets had overlaps ranging from 80.18% to 83.16%

with a mean of 81.44%. All correlation coefficients were

all positive and ranged from +0.78 to +0.86, and were all

significant (P<0.05). Cattle and wildebeest diets

overlaps ranged from 79.96% to 85.60% with a mean of

83.19%. All correlation coefficients were significantly

high (P<0.05) and ranged from +0.78 to +0.84. Kongoni

and wildebeest diet overlaps ranged from 83.06% to 88.49%

with a mean of 85.84%. Correlation coefficients were

significantly high (P<0.05) and ranged between +0.78, and

+ 0.83 .

Trophic diversity index measures niche breadth and

it complements studies on overlap and competition

(Hurtubia 1973). Generally, all animals in this study

had high trophic diversity indicating that they were not

dependent on any one particular plant species, but rather

selected from a wide spectrum of plants. Cattle had the

lowest trophic diversity (2.12 ± 0.06) and wildebeest had

the highest (2.20 ± 0.05), while kongoni's was

intermediate (2.16 ± 0.05). Overall, wildlife had a

higher trophic diversity than cattle, but the difference

was not statistically significant (p=0.05).

74

Table 29: Percent similarity indices (S.I) and

correlation coefficients (Rs) between cattle

and wildlife diets during the dry season.

Animal combination Period S . I Rs n t

Cattle Vs kongoni May 80.18 0.78 19 5.139*

Jun 80.21 0.86 17 6.444*

Jul 83.16 0.83 19 6.020*

Aug 82.22 0.79 16 4.877*

Mean 81.44 ± 1.292

Cattle Vs wildebeest May 79.96 0.84 19 6.403*

Jun 82.77 0.82 18 5.751*

Jul 85.60 0.81 18 5.622*

Aug 84.45 0.78 18 5.042*

Mean 83.19 ± 2.124

Kongoni Vs wildebeest May 83.06 0.78 20 5.331*

Jun 85.80 0.83 18 6.013*

Jul 88.49 0.81 19 5.836*

Aug 86.04 0.80 18 5.448*

Mean 85.84 ± 1.92

* - significant (P<0.05).

5.1.5 DiscussionsAll the study animal species almost had similar

plant varieties in their diets both during the wet and

dry season. This observation however contradicted the

results of Casebeer and Koss (1970) in the same general

75

area. They found that the diets of the animals had a

greater variety during the dry season than during the wet

season, this is most likely explained in two ways. First

this study was carried out in a closed system. This may

have resulted in animals being limited in their preferred

habitats and home ranges and therefore limited plant

varieties to select from. Secondly the high abundance of

forage in the ranch during the study period may have not

necessitated a shift in the diets of the animals.

During the two seasons in 1993 three grass species;

D. macroblephara, P. mezianum and T. triandra dominated

the diets of the three animals. The preference for these

species were probably by virtue of being perennial and

having greater standing biomass than other grass species.

Kibet (1984) has advanced similar reasons for selection

of D. macroblephara by cattle. During the wet season the

three most preferred plant species constituted about 70%,

61% and 63% for cattle, kongoni and wildebeest,

respectively. Cattle had the highest browse component in

their diet of about 4%, while kongoni and wildebeest

diets constituted 1% and 2%, respectively. During the

dry season, T. triandra, D. macroblephara and P. mezianum

made up about 60%, 54% and 55% of cattle, kongoni and

wildebeest diets, respectively. The browse component

during the dry season increased to about 8%, 4% and 4%

for cattle, kongoni and wildebeest respectively. The

results clearly show an increased use of browse in the

dry season. This is in general agreement with earlier

76

studies by Payne and MacFarlane (1963) and Kayongo Male

(1986) who found that browse component in cattle diets

increased as the dry season progressed. This results,

however contradict those of Wangoi and Hansen (1987) who

concluded that cattle are predominantly grassers and the

browse component of their diet was higher during the wet

season than in the dry season. Whereas cattle in this

study browsed more during the dry season, Wangoi (1984)

working in Rendile part of Kenya observed that cattle

browsed most during the wet season when browse shoots are

most abundant. This difference is most likely due to

differences in the vegetation in the two areas. Most of

the trees and shrubs in the Rendile area which is in

ecological zone six were reported (same study) to be

devoid of leaves in the dry season. In the study area

(GRL) most shrubs and trees retained their leaves long

into the dry season (Personal observation), thereby

offering browse to the animals. Owaga (1975) observed

that wildebeest browsed about 1-2% of browse during the

wet season and almost none, at other times. This was in

contrast to the findings of this study, where wildebeest

took more browse during the dry season but about the same

amount during the wet season. This difference is most

likely due to management. Whereas Owaga did her study

on animals whose movement was unrestricted, this study

was carried out in a closed system which did not allow

for migration of the animals. This observation can be

explained in four ways. During the wet season the

77

animals utilized habitats which had less browse material

than during the dry season. It seems clear therefore

that the locality of the animals in relation to plant

availability will affect the proportion of browse in the

diet. Secondly, during the wet season, the new growth

from all plants is high in crude protein, digestibility,

has less fibre, and forage is abundant in general. There

would be no nutritional advantage in animals walking from

one shrub/tree to another in search of browse, whereas

the grass at ground level was more abundant, and

accessible. The energy cost involved would outweigh the

benefits. Thirdly this can also be viewed as a resource

utilization strategy whereby these animals make maximum

use of the grass when still growing and high in nutrient

contents before suddenly declining in quality with

advancement of maturity. This may result in browse

preservations, which declines less in quality, for the

critical dry season. Fourth, it has also been reported

(Van Soest, 1982) that young growing browse plants may

contain high levels of secondary compounds (tannins,

sapponins, etc). Some of these compounds interfere with

the digestive mechanisms in the animal, among other

effects. It is possible then, that these animals could

simply be avoiding much of the browse growth which may

contain the compounds. These explanations are however

neither mutually exclusive nor exhaustive because diet

selection under range condition is a complex process

involving both animal related and plant related factors,

78

as ameliorated by the environment.

During the wet and dry seasons, the dietary overlaps

and Spearman Rank Correlation Coefficients were high

among all the animal species. The findings of this study

clearly indicate that cattle, kongoni and wildebeest most

likely compete with each other for forage resources. Do

overlaps indicate competition?. Not necessarily. Field,

(1972) documented that competition only occurs when the

resources being shared are limiting. Many researchers in

East Africa have addressed the question of resource

partitioning and coexistence of East African ungulates.

Some of the theories on how coexistence is achieved

include; spatial and temporal distribution, grazing at

different herb layers, grazing of different plant parts,

differences in mouth structure and differences in body

size. Thus, it is possible that the study animals,

despite having high dietary overlaps, may have had

differential selectivity for plant parts. Talbot and

Talbot (1963) pointed out that ecological separation was

achieved by spatial separation. They also mentioned that

diets of various plains ungulates were complementary and

non duplicative. According to the findings of this

study, many diets were duplicated and none were

complementary. Spatial and temporal separation was

exhibited partially by kongoni and wildebeest.

Bell (1969) observed that during the wet season all

the species in the Serengeti plains occupy the same plain

catena. At the GRL, the study animals concentrated on

79

those habitats where grass was kept short and or in

growing condition by trampling, mowing and grazing. The

short grass was preferred as it was leafy and nutritious.

Thus, in the wet season, the animals were limited by

quality of food rather than the quantity. There was

super abundance of forage resources during this study

period hence critical levels that trigger competition may

have not been attained. This may be one of the reasons

why dietary overlaps were high. In conclusion the

dietary overlaps were lower between cattle and wildlife

than among wild animals.

Diet preference index for the animals differed from

season to season, and from plant to plant. The animals

were more selective during the wet season. This could be

explained in various ways first, the presence of higher

standing biomass during the wet season gave the animals

an opportunity to select diets that they prefer most.

During the dry season forage species that were not

preferred by the animals during the wet season were also

eaten resulting in lower diet preference. Secondly

animals select plant parts that are more nutritious. It

is possible therefore that the animals selected for those

plant parts that were more nutritious during the dry

season from those plants that were not preferred during

the wet season. Finally different habitats had different

proportions of standing biomass of each plant species.

The seasonal changes in the habitat preferences of the

animals therefore exposed them to those plants that were

80

not equally accessible during the wet season. These

seasonal changes in habitat preference by the animals

were reflected by the changes in diet preferences.

The changes in the proportion of the standing

biomass and variation in species of the herbaceous

materials from habitat to habitat can be explained by the

changes in cover, elevation, soils and utilization by the

animals. Habitats on the ridgetops and on the higher

slopes generally had a lower standing biomass, than those

on the lower slopes and lowlying areas. Similarly

habitats which had higher cover generally had higher

standing biomass than those with lower cover.

Interestingly habitats on the ridged tops and on the

slopes had lower cover and received heavier animal

utilization especially during the wet season. Habitats

on the lower slopes had higher cover and were least

utilized by the animals. This resulted in accumulation

of dry matter in these habitats hence higher standing

biomass.

5.2 Animal density and habitat preferences 5.2.1 Number and distribution of animals

This work was conducted to determine the preference

of the study animals for the different habitat types

during the wet and dry seasons. During the study period

cattle were herded throughout. Cattle thus had no free

choice of habitat. During wet months (January, February,

March and April) they grazed mainly in the Themeda

81

grassland and Balanites tree grassland. During this

period there were only 250 heads of mainly dairy cattle1

in the Ranch. From April onwards cattle numbers were

increased to 2,500 as animals for fattening (Penta herd)

were purchased. Cattle bomas were located at hill tops,

resulting in heavy livestock pressures being exerted on

these areas. On daily basis from April onwards cattle

were always herded from their bomas at the hilltops to

the low-lying areas where dams have been constructed for

watering the animals. This resulted therefore in the

animals utilizing most vegetation types. As the dry

season progressed (July and August) they were herded into

the Acacia woodland and A. xanthophloea bushland.

5.2.2 Animal density and habitat preferenceThe results of the habitat preference study were

based on observations of the wild herbivores (kongoni and

wildebeest) over a period of eight months. The seasonal

habitat preference indices, percent frequencies and

relative densities of all the six habitat types for

kongoni and wildebeest are shown on Tables 30 and 31,

respectively, and illustrated in Figure 6 (A) and 6 (B) .

It is evident that during the wet season, for kongoni and

wildebeest, the highest and positive preference indices

occurred in the Balanites tree grassland habitat on the

higher slopes (0.44 and 0.50 for kongoni and wildebeest

respectively) . All other habitat types had negative

indigenous animals kept for milk production.

82

preference indices. Themeda grassland, Balanites -

Acacia and A. drepanolobium dwarf tree grasslands had

-0.82, -0.32 and -0.44 preference indices, respectively,

for kongoni and -1.00, -0.34 and -0.65, respectively, for

wildebeest. The Acacia woodland and A. xanthophloea

bushland were avoided completely by the two animal

species (Preference Index -1.00).

During the dry season the two animal species

shifted their habitat preferences to the Balanites -

Acacia grassland. Preference indices were 0.45 and 0.36

for kongoni and wildebeest respectively. A. xanthophloea

bushland was avoided completely by both animals. Themeda

grassland, and Acacia woodland had negative preference

indices close to -1. Their respective indices were -0.92

and -0.76 for kongoni and -0.97 and -0.98 for wildebeest.

A. drepanolobium dwarf tree grassland had lower but

negative preference indices. They were -0.21 and -0.23

for kongoni and wildebeest respectively, while Balanites

tree grassland had preference indices of -0.47 and -0.25

for kongoni and wildebeest respectively.

83

Table 30: Habitat utilization by the kongoni, showing

relative densities (R.D) per Km2, percentage

frequencies (P.F) and preference indices (P.I)

during the wet and dry seasons in 1993, at the

GRL, Athi River, Kenya.

HabitatWet Season Dry Season

R. D P F P. I R D P F P . I

Baltre2 15. 56 60 51 0 44 4 74 16 33 -0 47

Balacc3 4 82 18 74 - 0 32 13 27 45 72 0 4 5

Accdrf4 4 00 15 56 - 0 44 8 13 27 99 - 0 2 1

Thtgsl1 1 33 5 .19 - 0 82 0 .67 2 .30 -0 92

Accwdl5 0 00 0 00 - 1 00 2 . 22 7 . 66 -0 76

Accxan6 0 00 0

oo -1 oo 0

oo 0 . 00 -1 00

Total 25 71 100 00 29 .03 100

oo

1Themeda grassland, 2Balanites tree grassland, 3Balanites - Acacia tree grassland, 4A. drepanolobium dwarf tree grassland, 5Acacia woodland and 6A. xanthophloea bushland.

84

Table 31: Habitat utilization by the wildebeest, showing

relative density (R.D) per km2, percentage

frequencies (P.F) and preference indices (P.I)

during the wet and dry seasons in 1993, at the

GRL, Athi River, Kenya.

HabitatWet Season Dry Season

R. D P .F P. I R D P. F P I

Baltre2 26 15 69. 94 0 50 10 07 25 76 -0 25

Balacc3 7 36 19. 70 -0 34 17 .73 45 33 0 36

Accdrf4 3 88 10. 36 -0 65 10 .75 27 49 -0 23

Thtgsl1 0 .00 0. 00 -1 00 0 .33 0 85 -0 97

Accwdl5 0 .00 0. 00 -1 00 0 .22 0 .57 -0 98

Accxan6 0 oo 0. 00 -1 00 0 oo 0 .00 -1 00

Total 37 .39 100 .00 39 .11 100 .00

1Themeda grassland, 2Balanites tree grassland, 3Balanites - Acacia tree grassland, 4A. drepanolobium dwarf tree grassland, 5Acacia woodland and eA. xanthophloea bushland.

An analysis of variance (Table 32) showed that there

was no significant variation in habitat preference

between the two animals (kongoni and wildebeest). This

phenomenon did not vary with seasons, however there was a

significant variation in habitat preference between the

habitats (F5 5=150.49, df=5, p<0.05).

85

Pref

eren

ce In

dex

1.0 -i

x©■ Oc©oc©

o o 000>

1

1“ r-w -- - o 6 Z

1.OJW et season

iDry season

Themeda grassland Acacia dwarf tree grassland

Balanites tree grassland Acacia wooland

Balanites Acacia tree grassland Acacia xanthophloea bushland

Figure 6: Seasonal changes in habitat preference for kongoni (A) and wildebeest (B) for the six different habitat types at GRL, Athi River, Kenya. 1993.

86

^ a kle 32: Analysis of variance of mean habitat

preferences of kongoni and wildebeest.

Source DF SS MS F Value

CORRECTED TOTAL 23 6.369

a n i m a l 1 0.011 0.011 1.74ns

SEASON 1 0.018 0.018 2.83ns

HABITAT 5 4.909 0.981 150.49*

ANIMAL* HABITAT 5 0.052 0.010 1.60ns

ANIMAL* SEASON 1 0.001 0.001 0.25ns

S E AS ON * HABI TAT 5 1.343 0.268 4*41.18

- significant (p<0.05), ns - non significant

Apparently the distribution of the animals seem to

fc>e neither dependent on the animal species and habitat

type nor on animal species and season interactions. This

is supported by the non significant animal - habitat and

animal - season interactions in our model. Habitat

preference was influenced by the interaction between

seasons and the environment (F55(=41.18, df = 5, p<0.05) .

However, there was no apparent relationship between

animal density (for both kongoni and wildebeest) and the

standing biomass. Evidence for this is shown by a low

Pearson's correlation coefficient (r=0.04, p=0.81 for

kongoni and r=0.014, p=0.93 for wildebeest). The two

animal species mostly preferred Balanites tree and

Balanites - Acacia tree grasslands. Balanites - Acacia

87

woodland was the least preferred habitat while the A.

xanthophloea woodland was completely avoided as shown in

Table 33.

Table 33: Mean habitat preferences of kongoni and

wildebeest.

Habitat Mean Habitat Preference n

Baltre2 0.55a* 4

Balacc3 0.037a 4

Accdrf4 -0.383a- 4

Thtgsl1 -0.928b 4

Accwdl5 -0.934b 4

Accxan6 -1.00b 4

means followed by same letter are not significantly different at P<0.05. 1Themeda grassland, 2Balanites tree grassland, 3Balanites - Acacia tree grassland,4A. drepanolobium dwarf tree grassland, bAcacia woodland and 6A. xanthophloea bushland.

5.2.3 Discussions fc

Changes in habitat utilization by the two wild

herbivores reflects preferred patterns of feeding habits

and feeding strategies. The relative densities of

kongoni and wildebeest in the Balanites tree grassland,

at the upper slopes were higher during the wet season,

than on the low-lying areas. At the beginning of the dry

season in May, the higher density in the low-lying areas,

(Balanites - Acacia and A . drepanolobium dwarf tree

grassland) areas indicated that they were utilizing these

88

areas in preference to the short grass areas at the upper

slopes. At this time there was less green biomass at the

upper slopes than on the low-lying areas. Kongoni

appeared to use the low-lying areas to a greater extent

than wildebeest both during the wet and dry seasons.

Kongoni and wildebeest are selective grazers that try to

maximize the amount of grass leaf in their diets.

Kongoni are probably able to select grass leaves in tall

grass area better than wildebeest. Wildebeest maximize

grass leaf intake by feeding in the short grass areas

with high leaf to stem ratio by feeding on the preferred

short grass areas on the upper slopes. The two wild

herbivores appeared not to occupy the Acacia woodland and

A. xanthophloea bushland habitats.

Elements of the grazing succession on the ranch as

described by Bell (1971) for the Serengeti ecosystem,

were apparent on the ranch especially during the wet

season. Short grass areas that received the heaviest

cattle utilization were preferred by the kongoni and

wildebeest. The low-lying areas were lightly utilized

during the wet season. One reason for this lighter

utilization could have been related to the uneven

distribution of cattle grazing pressure, which was

highest on the upper slopes (personal observation).

Therefore, a grazing succession in the low lying areas of

tall grass could not occur. The Balanites tree grassland

habitat clearly is the most preferred by cattle, kongoni

and wildebeest in the ranch only during the wet season

89

while Balanites - Acacia tree grassland is the most

preferred during the dry season. The seasonal habitat

preferences by kongoni and wildebeest is probably an

important mechanism of survival and optimum utilization

of the available resources, which aims at reducing the

impact on the dry season habitats during the wet season

90

CHAPTER SIXCONCLUSIONS AND RECOMMENDATIONS

The feeding habits of cattle, kongoni and wildebeest

whether as single species or in combination with other

animal species on the same range, have been studied by-

several researchers in different environments. Some

findings concur, while others conflict. In the current

study the animal species (kongoni, wildebeest and cattle)

increased browse components in their diets by about 100%

during the dry season. The animals, however, selected

similar diets in terms of plant species during both

seasons. The locality of the animals in relation to

plant availability affects the diet preferences and the

proportions of browse and grass in the diet.

The Balanites - Acacia tree grassland is the best

available habitat. It is the habitat upon which kongoni

and wildebeest depend on for their survival in the ranch

during the dry season, when the forage resources are

limiting. Without this habitat these animals would lose

their body condition, which will eventually result in

increased mortalities and reduced population size.

Balanites tree grassland, however, is the most preferred

habitat but is only available during the wet season.

Generally the proportion of herbaceous standing

biomass, and cover of the woody species is dependent on

the interactions between the plants and the environment.

Diet preference is dependent on interactions between

91

season and plant and on interaction between animal and

plant. Therefore any changes of factors in the

interactions will result in changes in the dependable/

factor. One therefore can manipulate any one of these

habitat factors so as to achieve a desired objective.

Such habitat manipulation when being implemented should

take into consideration herbage production, habitat

requirements and the diet preference of the animals in

question. In a closed system animals are limited in

their habitats and plant variety to select diets from.

This therefore calls for proper management. This can be

achieved by harvesting/cropping the animals and through

habitat manipulation. Habitat manipulation should aim at

increasing heterogeneity rather than homogeneity in

habitats. This will allow the animals to maximize their

habitat and diet selection and hence result in increased

productivity.Finally it is recommended that similar studies be

undertaken on other animal species in the ranch. This

will enable comprehensive management package to be

developed. Such a package should take into consideration

habitat and feed requirements of all the herbivores in

the ranch while taking into consideration the management

objective.

92

REFERENCESAndere, D.K. 1981. Wildebeest (Connochaetes taurinus)

and its food supply in Amboseli Basin. Afr. J.Ecol. 19:239-250.

Arnold, G.W. 1966a. The special senses in grazing animals. I. Sight and dietary habits in sheep.Aust. J. Agric. Res. 17:521-529.

Arnold, G.W. 1966b. The special senses in grazing animals II: Smell, taste and touch and dietary habits in sheep. Aust. J. Agric. Res. 17:531-542.

Arnold, G.W. and M.C. Dudzinki. 1978. Ethology of free- ranging animals. Elselvier Sci. Publ. Co. Netherlands.

Bell, R.H.V. 1969. The use of the herb layer by grazing ungulates in Serengeti. In: A. Watson, ed. Animal population in relation to food resources. Symp.Brt. Ecol. Soc., Blackwell, Oxford, Edinborough.

Bell, R.H.V. 1971. A grazing ecosystem in Serengeti. Scientific Americana, 225:36-93.

Belovsky, G.E. and P.A. Jordan. 1978. Time-Energy budget of a moose. Theor. pop biol. 14:76-104

Ben-Shahar, R. 1991. Selectivity in large generalists herbivores; feeding patterns of African ungulates a semi-arid habitat. Afr. J. Ecol. 29:302-315.

Bjugstad, A.J., H.S. Crawford and I.N. Donald. 1970.Determining forage consumption by direct observation of domestic animals. In Range Wildl. Habitat Eval-A Res. Symp. US. Dep. Pub. No. 1147, 220p.(Quoted by Holechek et al. 1982b).

Bredon, R.M. and J. Wilson. 1963. The chemicalcomposition and nutritive value of grasses from semi arid area of Karamoja as related to ecology and types of soils. East Afr. Agric. For. J. 29:134- 142.

Bryan, D.G., M. Elfatih and J.L. Holechek. 1983.Mastication effect on cattle diets determined by microhistological analysis. J. Range Manage. 36:475-478.

Casebeer, R.L. and G. Koss. 1970. Food habits ofwildebeest, zebra, hartebeest and cattle in Kenya, Masailand. E. Afr. Wild. J. 8:25-36.

93

Cavender, R.B. and R.M. Hansen. 1970. The microscope method used for herbivore diet estimates and botanical analysis of litter and mulch at Pawnee site. Grassl. Biome, U.S. Int. Biol. Prog. Techn.Rep. No.18.

Chacon, E. and T. H. Stobbs. 1976. Influence ofprogressive defoliation of a grass sward on the eating behaviour of cattle. Aust. J. Agric. Res. 27:709-27.

Church, D.C. 1975. Digestive physiology and nutrition of ruminants. O & B books, Corvallis Ore.

Chippendale, G. 1962. Botanical composition of Kangaroo and cattle stomach contents. Aust. J. Sci. 25:21- 22.

Chamrad, A. D., and T.W. Box. 1964. A point frame for sampling rumen stomach contents. Aust. J. Sci. 25:21-22.

Cook, C.W. and L.A. Stoddart. 1953. The Quadry ofutilization and preference. J. Range Manage. 8:327-335.

Croker, B.H. 1959. A method for estimating thebotanical composition of the diet of sheep. N.Z.J. Agr. Res. 2:72-85. (Quoted by Holechek et al.1982b).

Demment, M.W., E.A. Laca and G.B. Greenwood. 1986.Intake in grazing ruminants: A conceptual framework, Feed, Intake Symposium, pp 208-225.

Dieter Mueller-Domboise and Heinz Ellenberg. 1974. Aims and methods of vegetation ecology. John Willey and sons inc. Dougall, H.W., V.M., V.M. Drysdale, and P.E. Glover. 1964. The chemical composition of Kenya browse and pasture herbage. E. Afr. Wildl. J. 2:86-11.

Duncan, P., T.J. Foose, I.J. Gordon, C.G. Gakahu and M. Lloyd. 1990. Comparative nutrient extraction by grazing bovids and equids: a test of the nutritional model of equid/bovid competition and coexistence. Oecologia, 84: 411 - 418.

Field, C.R. 1968. The food habits of some wildungulates in reaction to land use and management.E. Afr. Agric. For. J. 33:159-162.

Field, C.R. 1972. The food habits of wild ungulates in Uganda by analysis of stomach contents. E . Afr. Wildl. J. 10:17-42.

94

Field, C.R. 1975. Climate and food habits of ungulates on Galana Ranch. E. Afr. Wildl. J. 13:203-220.

Field, D.I. and J.K. Potere. 1972. The value of browseplants. 1n: Over-browsing of Capparis tomentosa bushes by goats in Ruwensori National Park, Uganda. Afr. J. Ecol. 18:11-17.

Foppe, F.M. 1984. Microhistological technique training program. Misc. Publ. Composition analysis lab range science dept. Colorado State University.

French, M.H. 1957. Nutritive value of tropical grasses and fodders. Herb Abst. 27:1-9.

Gakahu, C.G. 1982. Feeding strategies of plains Zebra, Equus quaga boehmii in Amboseli Ecosystem. PhD. Thesis, University of Nairobi.

Geist, V. 1974. On the relationship of social evolution and ecology in ungulates. Am. Nat. 14:205-220.

Githaiga, J.M. 1991. Grazing Speed: The test of a model. MSc. Thesis University of Nairobi.

Goatcher, W.D. and D.C. Church. 1970. Review of somenutritional aspects of the sense of taste. J. Anim. Sci. 31:973-981.

Gwynne, M.D. and R.H.V. Bell. 1968. Selection ofvegetation components by grazing ungulates in the Serengeti National Park. Nature. 220:390-393.

Hanley, T.A. 1982. The nutritional basis for foodselection by ungulates. J. Range Manage. 35:146- 151.

Hansen, R.M. 1971. Composition in wild sheep diets. First trans. N. Amer. Wild. Sheep Confr. ppl80.

Hansen, R.M., T.M. Foppe, M.B. Gilbert, R.C. Clark and H.W. Reynolds. 1984. The microhistological analyses of faeces as an estimator of herbivore dietary. Misc. Pub. Composition analysis lab.Range Science Department. Colorado State University.

Heady, H.F. 1975. "Range management". McGraw Hill book Co., New York.

Heady, H.F. and G.M. Van Dyne. 1965. Prediction of weight composition from point samples on clipped herbage. J. Range Manage. 18:144-149.

Hillman, C. and A.K.K. Hillman. 1977. Mortality ofwildlife in Nairobi National Park During the drought of 1973-1974. E. Afri. Wildl. J. 10:1-8.

95

Hillman, J.C. 1979. The biology of the Eland(Taurotragus oryx Pallas) in the wild. PhD. Thesis, University of Nairobi.

Hofmann, R.R. 1973. The Ruminant stomach. East African Monographs in Biology, volume 2. East African literature bureau.

Holechek, J.L., J.M. Varva, J.M. Skovlin and R.L.Philips. 1981. Diet quality and performance of cattle on grassland and forest range. J. Ani. Sci. 53:291-298.

Holechek, J.L. and B.D. Gross. 1982. Evaluation ofdifferent procedures for microhistological analysis. J. Range Manage. 35:541-542.

Holechek, J.L., M. Varva and R.D. Pieper. 1982a. Botanical composition determination of range herbivore diets: A review. J. Range Manage. 35:309-315.

Holechek, J.L., M. Varva, J. Skovlin, W.C. Krueger.1982b. Cattle diets in the Blue Mountains of Oregon II forests. J. Range Manage. 35:239.

Hopcraft, D. 1975. Productivity comparison betweenThompson's Gazelle and cattle, and their Relation to the Ecosystem in Kenya. PhD. Thesis, Cornel University.

Hoppe, P.P. 1977. Rumen fermentation and food selection in East African sheep, goats, Thomson's gazelle, Grant's Gazelle and Impala. J. agric. Sci., Camb. 89:129-135.

Hungate, R.E. 1966. The Rumen and its Microbes.Academic press, New York.

Hurtubia, J. 1973. Tropic diversity measurement insympatric predatory species. Ecology, 54:885-890.

Jarman, P.J. 1974. The Social organization of antelopein relation to their ecology. Behaviour, 48:215-67.

Jarman, P.J. and A.R.E. Sinclair. 1979. Feedingstrategy and pattern of resource partitioning in ungulates. In: Serengeti dynamics of an ecosystem. University of Chicago Press, Chicago.

Kayongo Male, H. 1986. Feeding habits, forage quality and food intake by Zebu cattle grazing natural rangeland in IPAL, study area. IPAL Technical Report No. E-8; 201.

96

Kerr, J.A. , V.J. Wilson and H.H. Roth. 1970. Studies on agricultural utilization of semi-domesticated Eland (Taurotragus oryx) in Rhodesia. Rhod. J.Agr. Res. 8:71-71 (Quoted by Ng'ethe and Box. 1976) .

Kibet, P.K. 1984. Influence of browse on cattle diets in Acacia savannah of East Africa. MSc. Thesis Texas A and M University, Texas.

Kilcher M.R. 1981. Plant development, stage ofmaturity, and nutrient composition. J. Range manage. 34:363-364.

Krebs, J.R. 1978. Rules for predators. In: Behavioural ecology: An Evolutionary Approach, pp 23 - 68. Eds.

Krueger, W.C., W.C. Lay Cock, and D .A . Price. 1974.Relationships of taste, smell, sight and touch to forage selection. J. Range Manage. 27:258-262.

Lamprey, H.F. 1963. Ecological separation of the large mammal species in Tarangire Game Reserve,Tanganyika. E. Afr. Wildl. J. 1:63-92.

Lawton, R.M. 1968. The value of browse in dry tropics. East Afr. agric. For. J. 33:227-231.

Lotus Development Corporation. 1991. Lotus Freelance Graphics, version 4.0. 55. Cambridge parkway,Cambridge.

MacDowell, R.E., D.G. Sister, E.C. Schemerhon, J.D. Reed and R.P. Bauer. 1988. Game or cattle for meat production on Kenya rangelands. Dept, of Anim.Sci., New York College of Agriculture and life sciences. Cornell University, Ithaca, New York.

Malechek, J.C. and F.D. Provenza. 1983. Feedingbehaviour and nutritional characteristics of goats on rangelands. World Animal Review, 47: P.38.

Mentis, M.T. and R.R. Duke. 1976. Carrying capacitiesof natural veld in Netal for forage wild herbivores. S. Afr. J. Wild. Res. 6:65-74.

Mentis, M.T. 1977. Stocking rates and carryingcapacities for ungulates on African rangelands. S. Afr. J. Wildl. res. 7:89-98.

Ng'ethe, J.C. and T.W. Box. 1976. Botanical composition of Eland and goat diets on an acacia-grassland community in Kenya. J. Range Manage. 29:290-293.

Oosting, J.H. 1956. The study of plant communities. W. H. Freeman and Company, San Francisco, California.

97

Owaga, M.L. 1975. The feeding ecology of the wildebeest and zebra in Athi Kapiti Plains. E . Afr. Wildl. J. 13:375-383.

Owen-Smith, N. 1982. Factors influencing theconsumption of plant products by large herbivores.In: Ecology of tropical Savannas. (Eds B.J. Huntley and B.H. Walker) PP. 359-404. Springer-verlag, New York.

Payne, W.J.A. and J.S. Macfarlane. 1963. A brief study of cattle browsing behaviour in a semi-arid area of Tanganyika. E. Afr. For. J. 29:131.

Pratt, D.J. and M.D. Gwynne. 1977. "Rangelandmanagement and ecology in East Africa". Robert E. Krueger Publ. Co. Huntington, New York.

Pratt, D.J., P.J. Greenway and M.D. Gwynne. 1966. Aclassification of East African Rangelands. J. App. Ecology, 3:369-392.

Sanders, K.D., B.E. and G. Scott. 1980. Bite count Vs fecal analysis for range animals diets. J. Range Manage. 32:146-149.

SAS Inc., 1987. SAS (r) Proprietary Software Release 6.04. Cary, NC 27512-8000, U.S.A.

* Scother, J.S.B. 1979. A review of fecal analysistechniques for determination of the diet of wild grazing herbivores. Proc. Grassld. Soc. Afr. 14:141-146. (Quoted by Holechek et al. 1982a)

Shannon, C.E. 1948. A mathematical theory ofcommunication. Bell Systems Techn. J. 27:378-423.

Sinclair, A.R.E. and M. Norton - Griffiths, 1979,Serengeti: Dynamics of an ecosystem. The University of Chicago press, Chicago and London.

Sinclair, A. R. E. 1972. Food selection and competition in East African Buffalo (Syncerus cafer Sparman).E. Afri. Wildl. J. 10: 77-78.

Slater, J. and R.J. Jonnes. 1971. Estimation of diets selected by grazing animals from microscopic analysis of faeces. J. Aust. Inst. Agri. Sci. 37:238. (Quoted by Holechek 1982a).

Snedecor, G.W. and W.G. Cochran. 1967. Statisticalmethods. The Iowa State University Press, Ames, Iowa, U.S.A.

98

* Sparks, D.R., and J.C. Malechek. 1968. Estimatingpercentage in diets using a microscopic technique.J. Range Manage. 21:264-265.

Stanley - Price, M.R. 1977. The estimation of food intake and its seasonal variation, in the hartebeest. E. Afr. Wildl. J. 15:107-124.

Stelfox, J.B. 1985. Mixed species Game Ranching on the Athi Kapiti plains, Kenya. PhD. Thesis. University of Albert: In: Sinnary, A.S. and J.J. Hebrard. 19 91. A new approach for detecting visibility bias. Afr. J. Ecol. 29:222-28.

Stewart, D.R.M. and J. Stewart. 1970. Food preference data analysis for african plains ungulates.Zoologica Africana, 5:115-129.

Stobbs, T.H. 1973. The effect of plant structure onintake of tropical pastures. Aust. J. Agric. Res. 24:809-19.

Stobbs, T.H. 1975. The effect of plant structure on the intake of Tropical pastures. Ill Influence of fertilizer nitrogen on the size of bite harvested by Jersey cows grazing Setaria anceps CV. Kazungula swards. Aust. J. Agric. Res. 26:997-1007.

Taerum, R. 1970. A note on the chemical contents ofsome East African grasses. East Afr. Agric. For. J. 46:171-176.

Talbot, L.M. and M.H. Talbot. 1962. Food preferences of some East African wild Ungulates. E. Afr. Agric. For. J. 27:131-138.

Talbot, L.M. and M.H. Talbot. 1963. The wildebeest in western Maasailand East Africa. Wildlife Monograph, 2:88.

Todd, J.W. and R.M. Hansen. 1973. Plant fragments in faeces of big horns as indicators of food habits.J. Wild. Manage. 37:363-367.

Van Dyne, G.M., N.R. Brockngton, Z. Szocs, J, Duek and C. A. Ribic. 1980. Large herbivore subsystem. In: Grasslands System Analysis and Man. (Edn A.I. Brey Meyer and G.M. Van Dyne) pp 269 - 537. I.B.P. Synthesis 19. Cambridge University Press,Cambridge.

Van Soest, P.J. 1982. Nutritional ecology of the ruminant. O&B Books Inc. Corvallis, Ore.

99

Van Soest, P.J. 1967. Development of a comprehensive system of feed analysis and its application to forage. J. Anim. Sci. 26: 119 - 128

Van Zyl, J.H.M. 1965. The vegetation of S.A. Lombard Nature Reserve and its utilization by certain antelopes. Zoologica Africana, 1:55. Quoted by Ng'ethe and Box 1976).

tVavra, M. and J.L. Holechek. 1980. Factors influencing microhistological analysis of herbivore diets. J. Range Manage. 33:371-374.

Vessey - Fitzgerald, D.F. 1960. Grazing successionamong East African game animals. J. Mammal. 41:161- 72.

Viljoen, P.J. 1989. Habitat selection and preferred food plants of a desert-dwelling elephant in the northern Namib Desert, South West Africa/Namimbia.J. Afr. Ecol. 27:227-240.

Wangoi, E. Migongo. 1984. The tropic relations ofdomestic animals in he central part of Rendile land in Northern Kenya. Ph.D Diss. Colorado State Univ. Fort Collins, Co.

Wangoi Migongo-Bake and R.M. Hansen. 1987. Seasonaldiets of camels, cattle, sheep and goats in a common range in East Africa. J. Range Manage. 40:76.

Welch, J.G. and A.M. Smith. 1969. Influence of forage quality on rumination time in sheep. J. Anim. Sci. 28:813-818.

* Wilson, A.E., S.M. Hirst and R.P. Ellis. 1977.Determination of feeding preferences in wild ruminants from fecal samples. J. Wild. Manage. 41:70-75.

100

APPENDIXScientific names and authorities for plants and animals species mentioned in the text

1. GrassesDigitaria macroblephara (Hack.) Stapf Penisetum mezianum Leeke Themeda triandra Fork Cynodon dactylon (L.) Pers.Harpachne schimperi A. Rich.Penisetum stramineum PeterBothriochloa inscupta (A. Rich) A. CamusPenisetum masaicum StapfLintonia nutans StapfIschaemum afrum (J. F. Gmel.) DandyPanicum maximum Jacq.Cenchrus ciliaris L.Chloris gayana Kunth.Heteropogon contortus (L.) Beauv. ex R. & S. Sporobolus pellucidus Hochst.

2 . TressAcacia xanthophloea Benth.Acacia drepanolobium Harms ex Sjostedt Acacia seyal Del.Acacia stuhlmanii Taub Balanites glabra Mildbr. & Schleht

3. AnimalsCommon nameZebraKongoniWildebeestThomsons' gazelleGrants' gazelleElandMaasai giraffe Impala Ostrich Catlle

Scientific names and authority Equus burchelli Gray Alcephalus buselaphus Pallas Connochaetes taurinus Burchell Gazella thomsoni Gunther Gazella granti Brooke Taurotrogus oryx Pallas Giraffa Camelopardalis L. Aepyceros melampus Lichtenstein Structhio camelus L.Bos indicus

3t

101